[ { "name": "mlx-community/Devstral-Small-2505-4bit", "provider": "mlx-community", "parameter_count": "3.68354B", "parameters_raw": 3683537920, "min_ram_gb": 3.1, "recommended_ram_gb": 4.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 66257, "hf_likes": 2, "release_date": "2025-05-21", "format": "mlx", "mlx_only": true, "collection": "Devstral Small 2505", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-0.6B-8bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.7, "recommended_ram_gb": 2.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 39254, "hf_likes": 7, "release_date": "2025-05-04", "format": "mlx", "mlx_only": true, "collection": "Qwen3", "description": "", "_discovered": true }, { "name": "mlx-community/whisper-large-v3-mlx", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "automatic-speech-recognition", "architecture": "", "hf_downloads": 38989, "hf_likes": 93, "release_date": "2026-04-12", "format": "mlx", "mlx_only": true, "collection": "Whisper", "description": "OpenAI Whisper speech recognition models in MLX format", "_discovered": true }, { "name": "mlx-community/gemma-3-27b-it-qat-4bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 16.5, "recommended_ram_gb": 20.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 34370, "hf_likes": 23, "release_date": "2025-04-19", "format": "mlx", "mlx_only": true, "collection": "Gemma 3 QAT", "description": "Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory.", "_discovered": true }, { "name": "mlx-community/diffusiongemma-26B-A4B-it-4bit", "provider": "mlx-community", "parameter_count": "26B", "parameters_raw": 26000000000, "min_ram_gb": 15.9, "recommended_ram_gb": 19.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 26614, "hf_likes": 32, "release_date": "2026-06-11", "format": "mlx", "mlx_only": true, "collection": "DiffusionGemma", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-0.6B-4bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 22696, "hf_likes": 14, "release_date": "2025-04-28", "format": "mlx", "mlx_only": true, "collection": "Qwen3", "description": "", "_discovered": true }, { "name": "mlx-community/Llama-3.3-70B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 41.2, "recommended_ram_gb": 49.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 14235, "hf_likes": 35, "release_date": "2024-12-06", "format": "mlx", "mlx_only": true, "collection": "Llama 3.3", "description": "", "_discovered": true }, { "name": "mlx-community/Kokoro-82M-bf16", "provider": "mlx-community", "parameter_count": "82M", "parameters_raw": 82000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 14162, "hf_likes": 56, "release_date": "2025-12-02", "format": "mlx", "mlx_only": true, "collection": "Kokoro TTS", "description": "Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers amazing quality.", "_discovered": true }, { "name": "mlx-community/Meta-Llama-3.1-8B-Instruct-bf16", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 19.4, "recommended_ram_gb": 23.6, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 11431, "hf_likes": 3, "release_date": "2024-10-19", "format": "mlx", "mlx_only": true, "collection": "Llama 3.1", "description": "", "_discovered": true }, { "name": "mlx-community/DeepSeek-V4-Flash-4bit", "provider": "mlx-community", "parameter_count": "284.333B", "parameters_raw": 284333146519, "min_ram_gb": 164.5, "recommended_ram_gb": 193.9, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 10061, "hf_likes": 23, "release_date": "2026-04-25", "format": "mlx", "mlx_only": true, "collection": "DeepSeek V4", "description": "", "_discovered": true }, { "name": "mlx-community/diffusiongemma-26B-A4B-it-8bit", "provider": "mlx-community", "parameter_count": "26B", "parameters_raw": 26000000000, "min_ram_gb": 30.9, "recommended_ram_gb": 37.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 9630, "hf_likes": 14, "release_date": "2026-06-10", "format": "mlx", "mlx_only": true, "collection": "DiffusionGemma", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-12b-coder-fable5-composer2.5-8bit", "provider": "mlx-community", "parameter_count": "12B", "parameters_raw": 12000000000, "min_ram_gb": 14.8, "recommended_ram_gb": 18.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 9322, "hf_likes": 21, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Gemma-4-12b-coder-fable5-composer2.5", "description": "MLX conversions of Gemma-4-12b-coder-fable5-composer2.5 for Apple Silicon Chips", "_discovered": true }, { "name": "mlx-community/Qwen3-Coder-30B-A3B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 18.2, "recommended_ram_gb": 22.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 8470, "hf_likes": 29, "release_date": "2025-08-06", "format": "mlx", "mlx_only": true, "collection": "Qwen3-Coder-MoE", "description": "💻 Significant Performance: among open models on Agentic Coding, Agentic Browser-Use, and other foundational coding tasks, achieving ~Claude Sonnet.", "_discovered": true }, { "name": "mlx-community/Llama-3.3-70B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 81.5, "recommended_ram_gb": 96.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 8078, "hf_likes": 15, "release_date": "2024-12-06", "format": "mlx", "mlx_only": true, "collection": "Llama 3.3", "description": "", "_discovered": true }, { "name": "mlx-community/Mistral-Nemo-Instruct-2407-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7675, "hf_likes": 15, "release_date": "2024-11-06", "format": "mlx", "mlx_only": true, "collection": "Mistral NeMo", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen1.5-0.5B-Chat-4bit", "provider": "mlx-community", "parameter_count": "500M", "parameters_raw": 500000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7458, "hf_likes": 4, "release_date": "2024-04-18", "format": "mlx", "mlx_only": true, "collection": "Qwen1.5", "description": "Qwen1.5 is the improved version of Qwen, the large language model series developed by Alibaba Cloud.", "_discovered": true }, { "name": "mlx-community/Meta-Llama-3.1-8B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7320, "hf_likes": 10, "release_date": "2024-11-26", "format": "mlx", "mlx_only": true, "collection": "Llama 3.1", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16-mlx-8Bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 63.1, "recommended_ram_gb": 74.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6670, "hf_likes": 19, "release_date": "2026-05-13", "format": "mlx", "mlx_only": true, "collection": "UNCENSORED Qwen 3.6 27B", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-Coder-480B-A35B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "480B", "parameters_raw": 480000000000, "min_ram_gb": 277.0, "recommended_ram_gb": 326.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6442, "hf_likes": 19, "release_date": "2025-07-22", "format": "mlx", "mlx_only": true, "collection": "Qwen3-Coder-MoE", "description": "💻 Significant Performance: among open models on Agentic Coding, Agentic Browser-Use, and other foundational coding tasks, achieving ~Claude Sonnet.", "_discovered": true }, { "name": "mlx-community/Qwen3-Next-80B-A3B-Thinking-4bit", "provider": "mlx-community", "parameter_count": "80B", "parameters_raw": 80000000000, "min_ram_gb": 47.0, "recommended_ram_gb": 56.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6159, "hf_likes": 4, "release_date": "2025-09-13", "format": "mlx", "mlx_only": true, "collection": "Qwen3 Next", "description": "Alibaba's first hybrid model, designed to cut resources and speed things up.", "_discovered": true }, { "name": "mlx-community/gemma-3n-E2B-it-lm-4bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6032, "hf_likes": 3, "release_date": "2025-06-29", "format": "mlx", "mlx_only": true, "collection": "Gemma 3n - Text Only (LM)", "description": "Google's Gemma 3n converted to MLX using mlx-lm", "_discovered": true }, { "name": "mlx-community/GLM-4.5-Air-8bit", "provider": "mlx-community", "parameter_count": "106.852B", "parameters_raw": 106852251264, "min_ram_gb": 123.9, "recommended_ram_gb": 146.3, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5830, "hf_likes": 9, "release_date": "2025-07-29", "format": "mlx", "mlx_only": true, "collection": "GLM-4.5-Air", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-Next-80B-A3B-Thinking-8bit", "provider": "mlx-community", "parameter_count": "80B", "parameters_raw": 80000000000, "min_ram_gb": 93.0, "recommended_ram_gb": 110.0, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5766, "hf_likes": 2, "release_date": "2025-09-13", "format": "mlx", "mlx_only": true, "collection": "Qwen3 Next", "description": "Alibaba's first hybrid model, designed to cut resources and speed things up.", "_discovered": true }, { "name": "mlx-community/gemma-4-12b-coder-fable5-composer2.5-4bit", "provider": "mlx-community", "parameter_count": "12B", "parameters_raw": 12000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5373, "hf_likes": 13, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Gemma-4-12b-coder-fable5-composer2.5", "description": "MLX conversions of Gemma-4-12b-coder-fable5-composer2.5 for Apple Silicon Chips", "_discovered": true }, { "name": "mlx-community/gemma-4-E2B-it-qat-4bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 4816, "hf_likes": 3, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Gemma 4 QAT", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-12B-it-qat-assistant-4bit", "provider": "mlx-community", "parameter_count": "12B", "parameters_raw": 12000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4787, "hf_likes": 2, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Gemma 4 MTP QAT", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-26B-A4B-it-assistant-bf16", "provider": "mlx-community", "parameter_count": "26B", "parameters_raw": 26000000000, "min_ram_gb": 60.8, "recommended_ram_gb": 72.2, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4256, "hf_likes": 19, "release_date": "2026-05-05", "format": "mlx", "mlx_only": true, "collection": "Gemma-4 Assistant (MTP)", "description": "", "_discovered": true }, { "name": "mlx-community/Meta-Llama-3-8B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4208, "hf_likes": 81, "release_date": "2024-04-19", "format": "mlx", "mlx_only": true, "collection": "Llama 3", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen2.5-Coder-14B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "14B", "parameters_raw": 14000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 3887, "hf_likes": 10, "release_date": "2024-11-11", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-Coder", "description": "Code-specific model series based on Qwen2.5", "_discovered": true }, { "name": "mlx-community/VibeThinker-3B-8bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 3882, "hf_likes": 3, "release_date": "2026-06-16", "format": "mlx", "mlx_only": true, "collection": "VibeThinker-3B", "description": "MLX conversions of VibeThinker-3B for Apple Silicon Chips. ", "_discovered": true }, { "name": "mlx-community/SmolLM3-3B-4bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 3745, "hf_likes": 6, "release_date": "2025-07-08", "format": "mlx", "mlx_only": true, "collection": "SmolLM3", "description": "", "_discovered": true }, { "name": "mlx-community/DeepSeek-R1-Distill-Qwen-7B-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 3435, "hf_likes": 22, "release_date": "2025-02-26", "format": "mlx", "mlx_only": true, "collection": "DeepSeek-R1-Distill", "description": "", "_discovered": true }, { "name": "mlx-community/Llama-4-Scout-17B-16E-Instruct-4bit", "provider": "mlx-community", "parameter_count": "17B", "parameters_raw": 17000000000, "min_ram_gb": 10.8, "recommended_ram_gb": 13.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 3343, "hf_likes": 10, "release_date": "2025-05-03", "format": "mlx", "mlx_only": true, "collection": "Llama 4", "description": "", "_discovered": true }, { "name": "mlx-community/diffusiongemma-26B-A4B-it-6bit", "provider": "mlx-community", "parameter_count": "26B", "parameters_raw": 26000000000, "min_ram_gb": 23.4, "recommended_ram_gb": 28.3, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 3314, "hf_likes": 4, "release_date": "2026-06-11", "format": "mlx", "mlx_only": true, "collection": "DiffusionGemma", "description": "", "_discovered": true }, { "name": "mlx-community/Ornith-1.0-35B-8bit", "provider": "mlx-community", "parameter_count": "35B", "parameters_raw": 35000000000, "min_ram_gb": 41.2, "recommended_ram_gb": 49.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 3288, "hf_likes": 4, "release_date": "2026-06-27", "format": "mlx", "mlx_only": true, "collection": "Ornith 1.0", "description": "MLX versions of Ornith 1.0", "_discovered": true }, { "name": "mlx-community/Ornith-1.0-35B-4bit", "provider": "mlx-community", "parameter_count": "35B", "parameters_raw": 35000000000, "min_ram_gb": 21.1, "recommended_ram_gb": 25.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 3243, "hf_likes": 5, "release_date": "2026-06-27", "format": "mlx", "mlx_only": true, "collection": "Ornith 1.0", "description": "MLX versions of Ornith 1.0", "_discovered": true }, { "name": "mlx-community/Qwen3.5-397B-A17B-4bit", "provider": "mlx-community", "parameter_count": "397B", "parameters_raw": 397000000000, "min_ram_gb": 229.3, "recommended_ram_gb": 270.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 3209, "hf_likes": 11, "release_date": "2026-02-18", "format": "mlx", "mlx_only": true, "collection": "Qwen-3.5", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-4b-it-8bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 3041, "hf_likes": 6, "release_date": "2025-03-19", "format": "mlx", "mlx_only": true, "collection": "Gemma 3", "description": "A collection of lightweight, state-of-the-art open models built from the same research and technology that powers the Gemini 2.0 models", "_discovered": true }, { "name": "mlx-community/Qwen3-ASR-0.6B-4bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2798, "hf_likes": 12, "release_date": "2026-01-29", "format": "mlx", "mlx_only": true, "collection": "Qwen3-ASR", "description": "This collection contains Qwen3-ASR & Qwen3-ForceAligner", "_discovered": true }, { "name": "mlx-community/gemma-4-31B-it-assistant-bf16", "provider": "mlx-community", "parameter_count": "31B", "parameters_raw": 31000000000, "min_ram_gb": 72.3, "recommended_ram_gb": 85.7, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2680, "hf_likes": 14, "release_date": "2026-05-05", "format": "mlx", "mlx_only": true, "collection": "Gemma-4 Assistant (MTP)", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-12b-coder-fable5-composer2.5", "provider": "mlx-community", "parameter_count": "12B", "parameters_raw": 12000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2499, "hf_likes": 2, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Gemma-4-12b-coder-fable5-composer2.5", "description": "MLX conversions of Gemma-4-12b-coder-fable5-composer2.5 for Apple Silicon Chips", "_discovered": true }, { "name": "mlx-community/GLM-5.1", "provider": "mlx-community", "parameter_count": "743.911B", "parameters_raw": 743911218432, "min_ram_gb": 428.7, "recommended_ram_gb": 504.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2493, "hf_likes": 5, "release_date": "2026-04-07", "format": "mlx", "mlx_only": true, "collection": "Glm 5.1", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3n-E4B-it-lm-4bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2438, "hf_likes": 8, "release_date": "2025-06-29", "format": "mlx", "mlx_only": true, "collection": "Gemma 3n - Text Only (LM)", "description": "Google's Gemma 3n converted to MLX using mlx-lm", "_discovered": true }, { "name": "mlx-community/gemma-4-31b-bf16", "provider": "mlx-community", "parameter_count": "31B", "parameters_raw": 31000000000, "min_ram_gb": 72.3, "recommended_ram_gb": 85.7, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 2399, "hf_likes": 25, "release_date": "2026-04-02", "format": "mlx", "mlx_only": true, "collection": "Gemma 4", "description": "", "_discovered": true }, { "name": "mlx-community/DeepSeek-R1-Distill-Llama-8B-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2380, "hf_likes": 11, "release_date": "2025-02-26", "format": "mlx", "mlx_only": true, "collection": "DeepSeek-R1-Distill", "description": "", "_discovered": true }, { "name": "mlx-community/GLM-5.1-MXFP4-Q8", "provider": "mlx-community", "parameter_count": "743.911B", "parameters_raw": 743911218432, "min_ram_gb": 428.7, "recommended_ram_gb": 504.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2187, "hf_likes": 4, "release_date": "2026-04-08", "format": "mlx", "mlx_only": true, "collection": "Glm 5.1", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-31b-8bit", "provider": "mlx-community", "parameter_count": "31B", "parameters_raw": 31000000000, "min_ram_gb": 36.6, "recommended_ram_gb": 43.9, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 2170, "hf_likes": 23, "release_date": "2026-04-02", "format": "mlx", "mlx_only": true, "collection": "Gemma 4", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16-mlx-fp16", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 63.1, "recommended_ram_gb": 74.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2120, "hf_likes": 2, "release_date": "2026-04-30", "format": "mlx", "mlx_only": true, "collection": "UNCENSORED Qwen 3.6 27B", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-E4B-it-assistant-bf16", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1900, "hf_likes": 11, "release_date": "2026-05-05", "format": "mlx", "mlx_only": true, "collection": "Gemma-4 Assistant (MTP)", "description": "", "_discovered": true }, { "name": "mlx-community/embeddinggemma-300m-4bit", "provider": "mlx-community", "parameter_count": "300M", "parameters_raw": 300000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "embedding", "capabilities": [ "mlx" ], "pipeline_tag": "sentence-similarity", "architecture": "", "hf_downloads": 1825, "hf_likes": 6, "release_date": "2025-09-04", "format": "mlx", "mlx_only": true, "collection": "EmbeddingGemma", "description": "", "_discovered": true }, { "name": "mlx-community/DeepSeek-R1-0528-Qwen3-8B-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1822, "hf_likes": 5, "release_date": "2025-05-30", "format": "mlx", "mlx_only": true, "collection": "DeepSeek R1 0528", "description": "", "_discovered": true }, { "name": "mlx-community/diffusiongemma-26B-A4B-it-5bit", "provider": "mlx-community", "parameter_count": "26B", "parameters_raw": 26000000000, "min_ram_gb": 19.7, "recommended_ram_gb": 23.9, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 1821, "hf_likes": 2, "release_date": "2026-06-11", "format": "mlx", "mlx_only": true, "collection": "DiffusionGemma", "description": "", "_discovered": true }, { "name": "mlx-community/GLM-5.1-DQ4plus-q8", "provider": "mlx-community", "parameter_count": "743.911B", "parameters_raw": 743911218432, "min_ram_gb": 428.7, "recommended_ram_gb": 504.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1752, "hf_likes": 6, "release_date": "2026-04-12", "format": "mlx", "mlx_only": true, "collection": "Glm 5.1", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-270m-it-8bit", "provider": "mlx-community", "parameter_count": "270M", "parameters_raw": 270000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1701, "hf_likes": 2, "release_date": "2025-08-09", "format": "mlx", "mlx_only": true, "collection": "Gemma 3-270m", "description": "", "_discovered": true }, { "name": "mlx-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-4bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 18.2, "recommended_ram_gb": 22.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 1647, "hf_likes": 1, "release_date": "2026-04-29", "format": "mlx", "mlx_only": true, "collection": "Nvidia Nemotron-3-Nano-Omni", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-TTS-12Hz-0.6B-Base-8bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.7, "recommended_ram_gb": 2.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 1631, "hf_likes": 4, "release_date": "2026-01-25", "format": "mlx", "mlx_only": true, "collection": "Qwen3-TTS", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-270m-it-4bit", "provider": "mlx-community", "parameter_count": "270M", "parameters_raw": 270000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1606, "hf_likes": 10, "release_date": "2025-08-14", "format": "mlx", "mlx_only": true, "collection": "Gemma 3-270m", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen2.5-VL-72B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 42.4, "recommended_ram_gb": 50.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 1572, "hf_likes": 8, "release_date": "2025-02-25", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-VL", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-Coder-30B-A3B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 35.5, "recommended_ram_gb": 42.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1566, "hf_likes": 6, "release_date": "2025-07-31", "format": "mlx", "mlx_only": true, "collection": "Qwen3-Coder-MoE", "description": "💻 Significant Performance: among open models on Agentic Coding, Agentic Browser-Use, and other foundational coding tasks, achieving ~Claude Sonnet.", "_discovered": true }, { "name": "mlx-community/Qwen2.5-Coder-3B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1521, "hf_likes": 3, "release_date": "2024-11-11", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-Coder", "description": "Code-specific model series based on Qwen2.5", "_discovered": true }, { "name": "mlx-community/Ministral-8B-Instruct-2410-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1499, "hf_likes": 12, "release_date": "2024-10-17", "format": "mlx", "mlx_only": true, "collection": "Ministral", "description": "", "_discovered": true }, { "name": "mlx-community/Llama-3.2-11B-Vision-Instruct-8bit", "provider": "mlx-community", "parameter_count": "11B", "parameters_raw": 11000000000, "min_ram_gb": 13.6, "recommended_ram_gb": 16.9, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 1330, "hf_likes": 9, "release_date": "2024-10-18", "format": "mlx", "mlx_only": true, "collection": "Llama 3.2", "description": "Meta goes small with Llama3.2, both text only 1B and 3B, and the 11B Vision models.", "_discovered": true }, { "name": "mlx-community/Ornith-1.0-35B-6bit", "provider": "mlx-community", "parameter_count": "35B", "parameters_raw": 35000000000, "min_ram_gb": 31.2, "recommended_ram_gb": 37.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 1281, "hf_likes": 3, "release_date": "2026-06-27", "format": "mlx", "mlx_only": true, "collection": "Ornith 1.0", "description": "MLX versions of Ornith 1.0", "_discovered": true }, { "name": "mlx-community/Step-3.5-Flash-4bit", "provider": "mlx-community", "parameter_count": "196.956B", "parameters_raw": 196956118272, "min_ram_gb": 114.2, "recommended_ram_gb": 134.9, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1272, "hf_likes": 10, "release_date": "2026-02-04", "format": "mlx", "mlx_only": true, "collection": "Step 3.5 Flash", "description": "By StepFun", "_discovered": true }, { "name": "mlx-community/Phi-3-mini-4k-instruct-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1269, "hf_likes": 12, "release_date": "2024-07-11", "format": "mlx", "mlx_only": true, "collection": "Phi-3", "description": "", "_discovered": true }, { "name": "mlx-community/Lance-3B-Video-bf16", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-video", "architecture": "", "hf_downloads": 1138, "hf_likes": 10, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Lance MLX", "description": "Feature-complete MLX port of ByteDance Lance: t2i, image_edit, x2t_image, t2v, video_edit, x2t_video.", "_discovered": true }, { "name": "mlx-community/MiMo-V2.5-ASR-MLX-8bit", "provider": "mlx-community", "parameter_count": "8.01859B", "parameters_raw": 8018587648, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "automatic-speech-recognition", "architecture": "", "hf_downloads": 1090, "hf_likes": 5, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "MiMo-V2.5-ASR", "description": "by Xiaomi, converted to MLX", "_discovered": true }, { "name": "mlx-community/Codestral-22B-v0.1-4bit", "provider": "mlx-community", "parameter_count": "22B", "parameters_raw": 22000000000, "min_ram_gb": 13.6, "recommended_ram_gb": 16.9, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1088, "hf_likes": 13, "release_date": "2024-05-29", "format": "mlx", "mlx_only": true, "collection": "Mistral (Mamba) Codestral", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-ASR-0.6B-8bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.7, "recommended_ram_gb": 2.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1076, "hf_likes": 3, "release_date": "2026-01-29", "format": "mlx", "mlx_only": true, "collection": "Qwen3-ASR", "description": "This collection contains Qwen3-ASR & Qwen3-ForceAligner", "_discovered": true }, { "name": "mlx-community/gemma-4-12B-it-qat-assistant-8bit", "provider": "mlx-community", "parameter_count": "12B", "parameters_raw": 12000000000, "min_ram_gb": 14.8, "recommended_ram_gb": 18.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1075, "hf_likes": 0, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Gemma 4 MTP QAT", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3n-E4B-it-4bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 1071, "hf_likes": 7, "release_date": "2025-07-13", "format": "mlx", "mlx_only": true, "collection": "Gemma 3n", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-0.6B-bf16", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 2.4, "recommended_ram_gb": 3.6, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1032, "hf_likes": 5, "release_date": "2025-04-28", "format": "mlx", "mlx_only": true, "collection": "Qwen3", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-VL-4B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 1030, "hf_likes": 3, "release_date": "2025-10-14", "format": "mlx", "mlx_only": true, "collection": "Qwen3-VL", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen2.5-72B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 42.4, "recommended_ram_gb": 50.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1022, "hf_likes": 7, "release_date": "2024-09-18", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5", "description": "The Qwen 2.5 models are a series of AI models trained on 18 trillion tokens, supporting 29 languages and offering advanced features such as instructio", "_discovered": true }, { "name": "mlx-community/medgemma-4b-it-4bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 1011, "hf_likes": 3, "release_date": "2025-06-09", "format": "mlx", "mlx_only": true, "collection": "MedGemma", "description": "Collection of Gemma 3 variants for performance on medical text and image comprehension to accelerate building healthcare-based AI applications.", "_discovered": true }, { "name": "mlx-community/GLM-4.6V-Flash-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 988, "hf_likes": 7, "release_date": "2025-12-08", "format": "mlx", "mlx_only": true, "collection": "GLM-4.6V", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen2.5-Coder-32B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 37.8, "recommended_ram_gb": 45.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 936, "hf_likes": 14, "release_date": "2024-11-11", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-Coder", "description": "Code-specific model series based on Qwen2.5", "_discovered": true }, { "name": "mlx-community/Nemotron-Cascade-2-30B-A3B-4bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 18.2, "recommended_ram_gb": 22.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 916, "hf_likes": 20, "release_date": "2026-03-20", "format": "mlx", "mlx_only": true, "collection": "Nemotron-Cascade 2", "description": "", "_discovered": true }, { "name": "mlx-community/GLM-4.5-Air-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 900, "hf_likes": 28, "release_date": "2025-07-28", "format": "mlx", "mlx_only": true, "collection": "GLM-4.5-Air", "description": "", "_discovered": true }, { "name": "mlx-community/DeepSeek-V4-Flash-6bit", "provider": "mlx-community", "parameter_count": "284.333B", "parameters_raw": 284333146519, "min_ram_gb": 246.2, "recommended_ram_gb": 289.9, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 897, "hf_likes": 2, "release_date": "2026-04-25", "format": "mlx", "mlx_only": true, "collection": "DeepSeek V4", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3.5-4B-MTP-4bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 881, "hf_likes": 1, "release_date": "2026-06-01", "format": "mlx", "mlx_only": true, "collection": "Qwen 3.x MTP", "description": "MLX MTP drafter checkpoints for Qwen 3.x speculative decoding with mlx-vlm.", "_discovered": true }, { "name": "mlx-community/VibeThinker-3B-4bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 876, "hf_likes": 0, "release_date": "2026-06-16", "format": "mlx", "mlx_only": true, "collection": "VibeThinker-3B", "description": "MLX conversions of VibeThinker-3B for Apple Silicon Chips. ", "_discovered": true }, { "name": "mlx-community/Mistral-Small-24B-Instruct-2501-4bit", "provider": "mlx-community", "parameter_count": "24B", "parameters_raw": 24000000000, "min_ram_gb": 14.8, "recommended_ram_gb": 18.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 848, "hf_likes": 14, "release_date": "2025-01-30", "format": "mlx", "mlx_only": true, "collection": "Mistral Small", "description": "", "_discovered": true }, { "name": "mlx-community/deepseek-r1-distill-qwen-1.5b", "provider": "mlx-community", "parameter_count": "1.5B", "parameters_raw": 1500000000, "min_ram_gb": 1.9, "recommended_ram_gb": 3.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 831, "hf_likes": 24, "release_date": "2025-02-26", "format": "mlx", "mlx_only": true, "collection": "DeepSeek-R1-Distill", "description": "", "_discovered": true }, { "name": "mlx-community/MiniMax-M2.7-8bit", "provider": "mlx-community", "parameter_count": "228.69B", "parameters_raw": 228689748992, "min_ram_gb": 264.0, "recommended_ram_gb": 310.7, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 817, "hf_likes": 1, "release_date": "2026-04-12", "format": "mlx", "mlx_only": true, "collection": "MiniMax-M2.7", "description": "", "_discovered": true }, { "name": "mlx-community/GLM-4.6-4bit", "provider": "mlx-community", "parameter_count": "352.798B", "parameters_raw": 352797829024, "min_ram_gb": 203.9, "recommended_ram_gb": 240.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 809, "hf_likes": 15, "release_date": "2025-09-30", "format": "mlx", "mlx_only": true, "collection": "GLM-4.6", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-4b-it-4bit-DWQ", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 791, "hf_likes": 1, "release_date": "2025-05-14", "format": "mlx", "mlx_only": true, "collection": "Gemma 3 DWQ", "description": "Gemma 3 distilled weight quantized (DWQ) models", "_discovered": true }, { "name": "mlx-community/embeddinggemma-300m-6bit", "provider": "mlx-community", "parameter_count": "300M", "parameters_raw": 300000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "embedding", "capabilities": [ "mlx" ], "pipeline_tag": "sentence-similarity", "architecture": "", "hf_downloads": 785, "hf_likes": 2, "release_date": "2025-09-04", "format": "mlx", "mlx_only": true, "collection": "EmbeddingGemma", "description": "", "_discovered": true }, { "name": "mlx-community/Hy-MT2-1.8B-4bit", "provider": "mlx-community", "parameter_count": "1.8B", "parameters_raw": 1800000000, "min_ram_gb": 2.0, "recommended_ram_gb": 3.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "translation", "architecture": "", "hf_downloads": 781, "hf_likes": 2, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "Hy-MT2", "description": "MLX conversions of Tencent Hy-MT2 1.8B and 7B.", "_discovered": true }, { "name": "mlx-community/DeepSeek-V4-Flash-bf16", "provider": "mlx-community", "parameter_count": "284.333B", "parameters_raw": 284333146519, "min_ram_gb": 655.0, "recommended_ram_gb": 769.7, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 780, "hf_likes": 1, "release_date": "2026-04-25", "format": "mlx", "mlx_only": true, "collection": "DeepSeek V4", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-E2B-it-assistant-bf16", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 757, "hf_likes": 6, "release_date": "2026-05-05", "format": "mlx", "mlx_only": true, "collection": "Gemma-4 Assistant (MTP)", "description": "", "_discovered": true }, { "name": "mlx-community/MiniMax-M2.7-4bit", "provider": "mlx-community", "parameter_count": "228.69B", "parameters_raw": 228689748992, "min_ram_gb": 132.5, "recommended_ram_gb": 156.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 742, "hf_likes": 2, "release_date": "2026-04-12", "format": "mlx", "mlx_only": true, "collection": "MiniMax-M2.7", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-1.7B-4bit-DWQ-053125", "provider": "mlx-community", "parameter_count": "1.7B", "parameters_raw": 1700000000, "min_ram_gb": 2.0, "recommended_ram_gb": 3.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 717, "hf_likes": 2, "release_date": "2025-06-01", "format": "mlx", "mlx_only": true, "collection": "Qwen3 DWQ Quants", "description": "High-quality 4-bit quants of the Qwen3 model family.", "_discovered": true }, { "name": "mlx-community/Qwen3-TTS-12Hz-0.6B-Base-4bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 715, "hf_likes": 9, "release_date": "2026-01-25", "format": "mlx", "mlx_only": true, "collection": "Qwen3-TTS", "description": "", "_discovered": true }, { "name": "mlx-community/Kimi-VL-A3B-Thinking-4bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 698, "hf_likes": 10, "release_date": "2026-01-27", "format": "mlx", "mlx_only": true, "collection": "Kimi-VL Thinking", "description": "", "_discovered": true }, { "name": "mlx-community/chatterbox-turbo-fp16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 694, "hf_likes": 23, "release_date": "2025-12-17", "format": "mlx", "mlx_only": true, "collection": "Chatterbox TTS", "description": "Chatterbox and Chatterbox Turbo By ResembleAI", "_discovered": true }, { "name": "mlx-community/Llama-3.2-11B-Vision-Instruct-abliterated", "provider": "mlx-community", "parameter_count": "11B", "parameters_raw": 11000000000, "min_ram_gb": 7.3, "recommended_ram_gb": 9.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 694, "hf_likes": 7, "release_date": "2024-12-16", "format": "mlx", "mlx_only": true, "collection": "Llama 3.2", "description": "Meta goes small with Llama3.2, both text only 1B and 3B, and the 11B Vision models.", "_discovered": true }, { "name": "mlx-community/embeddinggemma-300m-8bit", "provider": "mlx-community", "parameter_count": "300M", "parameters_raw": 300000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "embedding", "capabilities": [ "mlx" ], "pipeline_tag": "sentence-similarity", "architecture": "", "hf_downloads": 691, "hf_likes": 5, "release_date": "2025-09-04", "format": "mlx", "mlx_only": true, "collection": "EmbeddingGemma", "description": "", "_discovered": true }, { "name": "mlx-community/medgemma-1.5-4b-it-bf16", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 657, "hf_likes": 8, "release_date": "2026-01-14", "format": "mlx", "mlx_only": true, "collection": "MedGemma-1.5", "description": "MedGemma-1.5 models in MLX format. See original repo: https://huggingface.co/google/medgemma-1.5-4b-it", "_discovered": true }, { "name": "mlx-community/Llama-3.3-70B-Instruct-3bit", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 31.2, "recommended_ram_gb": 37.4, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 637, "hf_likes": 8, "release_date": "2024-12-06", "format": "mlx", "mlx_only": true, "collection": "Llama 3.3", "description": "", "_discovered": true }, { "name": "mlx-community/MiniCPM-V-4.6-bf16", "provider": "mlx-community", "parameter_count": "1.30043B", "parameters_raw": 1300428016, "min_ram_gb": 4.0, "recommended_ram_gb": 5.5, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 629, "hf_likes": 3, "release_date": "2026-05-13", "format": "mlx", "mlx_only": true, "collection": "MiniCPM-V 4.6", "description": "MLX variants of MiniCPM-V 4.6, 1.3B parameters (SigLIP2 400M vision encoder + Qwen3.5-0.8B LLM), repo: https://huggingface.co/openbmb/MiniCPM-V-4.6", "_discovered": true }, { "name": "mlx-community/Kokoro-82M-4bit", "provider": "mlx-community", "parameter_count": "82M", "parameters_raw": 82000000, "min_ram_gb": 1.0, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 618, "hf_likes": 8, "release_date": "2026-01-05", "format": "mlx", "mlx_only": true, "collection": "Kokoro TTS", "description": "Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers amazing quality.", "_discovered": true }, { "name": "mlx-community/MiMo-V2.5-ASR-MLX", "provider": "mlx-community", "parameter_count": "1.25344B", "parameters_raw": 1253440000, "min_ram_gb": 1.7, "recommended_ram_gb": 2.8, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "automatic-speech-recognition", "architecture": "", "hf_downloads": 595, "hf_likes": 1, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "MiMo-V2.5-ASR", "description": "by Xiaomi, converted to MLX", "_discovered": true }, { "name": "mlx-community/Phi-3-mini-128k-instruct-4bit", "provider": "mlx-community", "parameter_count": "597.212M", "parameters_raw": 597212160, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 594, "hf_likes": 15, "release_date": "2024-07-11", "format": "mlx", "mlx_only": true, "collection": "Phi-3", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3n-E2B-it-4bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 590, "hf_likes": 10, "release_date": "2025-07-12", "format": "mlx", "mlx_only": true, "collection": "Gemma 3n", "description": "", "_discovered": true }, { "name": "mlx-community/MiniCPM-V-4.6-4bit", "provider": "mlx-community", "parameter_count": "1.04032B", "parameters_raw": 1040315632, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 588, "hf_likes": 3, "release_date": "2026-05-13", "format": "mlx", "mlx_only": true, "collection": "MiniCPM-V 4.6", "description": "MLX variants of MiniCPM-V 4.6, 1.3B parameters (SigLIP2 400M vision encoder + Qwen3.5-0.8B LLM), repo: https://huggingface.co/openbmb/MiniCPM-V-4.6", "_discovered": true }, { "name": "mlx-community/jinaai-ReaderLM-v2", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 579, "hf_likes": 25, "release_date": "2025-01-17", "format": "mlx", "mlx_only": true, "collection": "Jina Reader-LM", "description": "Convert HTML content to LLM-friendly Markdown/JSON content", "_discovered": true }, { "name": "mlx-community/Llama-4-Maverick-17B-16E-Instruct-4bit", "provider": "mlx-community", "parameter_count": "17B", "parameters_raw": 17000000000, "min_ram_gb": 10.8, "recommended_ram_gb": 13.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 530, "hf_likes": 7, "release_date": "2025-04-06", "format": "mlx", "mlx_only": true, "collection": "Llama 4", "description": "", "_discovered": true }, { "name": "mlx-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-8bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 35.5, "recommended_ram_gb": 42.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 499, "hf_likes": 1, "release_date": "2026-04-29", "format": "mlx", "mlx_only": true, "collection": "Nvidia Nemotron-3-Nano-Omni", "description": "", "_discovered": true }, { "name": "mlx-community/Mistral-7B-Instruct-v0.2", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 492, "hf_likes": 20, "release_date": "2023-12-23", "format": "mlx", "mlx_only": true, "collection": "Mistral", "description": "", "_discovered": true }, { "name": "mlx-community/DeepSeek-R1-0528-4bit", "provider": "mlx-community", "parameter_count": "104.939B", "parameters_raw": 104938540544, "min_ram_gb": 61.3, "recommended_ram_gb": 72.8, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 471, "hf_likes": 17, "release_date": "2025-05-29", "format": "mlx", "mlx_only": true, "collection": "DeepSeek R1 0528", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-12B-it-qat-assistant-6bit", "provider": "mlx-community", "parameter_count": "12B", "parameters_raw": 12000000000, "min_ram_gb": 11.3, "recommended_ram_gb": 14.2, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 467, "hf_likes": 0, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Gemma 4 MTP QAT", "description": "", "_discovered": true }, { "name": "mlx-community/Lance-3B-bf16", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 466, "hf_likes": 8, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Lance MLX", "description": "Feature-complete MLX port of ByteDance Lance: t2i, image_edit, x2t_image, t2v, video_edit, x2t_video.", "_discovered": true }, { "name": "mlx-community/Hy-MT2-7B-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "translation", "architecture": "", "hf_downloads": 462, "hf_likes": 1, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "Hy-MT2", "description": "MLX conversions of Tencent Hy-MT2 1.8B and 7B.", "_discovered": true }, { "name": "mlx-community/Ornith-1.0-35B-5bit", "provider": "mlx-community", "parameter_count": "35B", "parameters_raw": 35000000000, "min_ram_gb": 26.2, "recommended_ram_gb": 31.5, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 461, "hf_likes": 0, "release_date": "2026-06-27", "format": "mlx", "mlx_only": true, "collection": "Ornith 1.0", "description": "MLX versions of Ornith 1.0", "_discovered": true }, { "name": "mlx-community/granite-4.1-30b-mxfp8", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 18.2, "recommended_ram_gb": 22.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 445, "hf_likes": 2, "release_date": "2026-04-30", "format": "mlx", "mlx_only": true, "collection": "Granite 4.1", "description": "By IBM", "_discovered": true }, { "name": "mlx-community/gemma-4-12B-coder-fable5-composer2.5-v1-4bit-msq", "provider": "mlx-community", "parameter_count": "12B", "parameters_raw": 12000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 436, "hf_likes": 6, "release_date": "2026-06-18", "format": "mlx", "mlx_only": true, "collection": "Gemma-4-12b-coder-fable5-composer2.5", "description": "MLX conversions of Gemma-4-12b-coder-fable5-composer2.5 for Apple Silicon Chips", "_discovered": true }, { "name": "mlx-community/gemma-3-1b-it-4bit-DWQ", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 429, "hf_likes": 0, "release_date": "2025-05-14", "format": "mlx", "mlx_only": true, "collection": "Gemma 3 DWQ", "description": "Gemma 3 distilled weight quantized (DWQ) models", "_discovered": true }, { "name": "mlx-community/Qwen2.5-72B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 83.8, "recommended_ram_gb": 99.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 427, "hf_likes": 3, "release_date": "2024-09-19", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5", "description": "The Qwen 2.5 models are a series of AI models trained on 18 trillion tokens, supporting 29 languages and offering advanced features such as instructio", "_discovered": true }, { "name": "mlx-community/Qwen2.5-Coder-14B-Instruct-bf16", "provider": "mlx-community", "parameter_count": "14B", "parameters_raw": 14000000000, "min_ram_gb": 33.2, "recommended_ram_gb": 39.8, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 424, "hf_likes": 2, "release_date": "2024-11-11", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-Coder", "description": "Code-specific model series based on Qwen2.5", "_discovered": true }, { "name": "mlx-community/gemma-4-12B-it-qat-assistant-5bit", "provider": "mlx-community", "parameter_count": "12B", "parameters_raw": 12000000000, "min_ram_gb": 9.6, "recommended_ram_gb": 12.1, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 424, "hf_likes": 0, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Gemma 4 MTP QAT", "description": "", "_discovered": true }, { "name": "mlx-community/granite-4.1-30b-mxfp4", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 18.2, "recommended_ram_gb": 22.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 418, "hf_likes": 2, "release_date": "2026-04-30", "format": "mlx", "mlx_only": true, "collection": "Granite 4.1", "description": "By IBM", "_discovered": true }, { "name": "mlx-community/Qwen2.5-7B-Instruct-1M-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 417, "hf_likes": 11, "release_date": "2025-01-26", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-1M", "description": "", "_discovered": true }, { "name": "mlx-community/DeepSeek-V4-Flash-5bit", "provider": "mlx-community", "parameter_count": "284.333B", "parameters_raw": 284333146519, "min_ram_gb": 205.4, "recommended_ram_gb": 241.9, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 415, "hf_likes": 0, "release_date": "2026-04-25", "format": "mlx", "mlx_only": true, "collection": "DeepSeek V4", "description": "", "_discovered": true }, { "name": "mlx-community/LFM2-8B-A1B-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 408, "hf_likes": 10, "release_date": "2025-10-08", "format": "mlx", "mlx_only": true, "collection": "💧LFM2-8B-A1B-MoE", "description": "Best in Class MoE, better than Qwen3. Optimised for Smaller devices sub 16 GB (M1/2/3/4) Apple Silicon.", "_discovered": true }, { "name": "mlx-community/Kokoro-82M-8bit", "provider": "mlx-community", "parameter_count": "82M", "parameters_raw": 82000000, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 400, "hf_likes": 8, "release_date": "2026-01-05", "format": "mlx", "mlx_only": true, "collection": "Kokoro TTS", "description": "Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers amazing quality.", "_discovered": true }, { "name": "mlx-community/granite-4.1-30b-nvfp4", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 18.2, "recommended_ram_gb": 22.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 397, "hf_likes": 2, "release_date": "2026-04-30", "format": "mlx", "mlx_only": true, "collection": "Granite 4.1", "description": "By IBM", "_discovered": true }, { "name": "mlx-community/Mellum-4b-base-4bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 389, "hf_likes": 3, "release_date": "2025-06-28", "format": "mlx", "mlx_only": true, "collection": "JetBrains Mellum", "description": "Series of code models by JetBrains", "_discovered": true }, { "name": "mlx-community/Qwen3-14B-4bit-DWQ-053125", "provider": "mlx-community", "parameter_count": "14B", "parameters_raw": 14000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 388, "hf_likes": 7, "release_date": "2025-06-02", "format": "mlx", "mlx_only": true, "collection": "Qwen3 DWQ Quants", "description": "High-quality 4-bit quants of the Qwen3 model family.", "_discovered": true }, { "name": "mlx-community/Hy3-preview-4bit", "provider": "mlx-community", "parameter_count": "295.034B", "parameters_raw": 295033528320, "min_ram_gb": 170.6, "recommended_ram_gb": 201.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 386, "hf_likes": 2, "release_date": "2026-04-27", "format": "mlx", "mlx_only": true, "collection": "Hy3 preview", "description": "By Tencent", "_discovered": true }, { "name": "mlx-community/MiniMax-M2-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 384, "hf_likes": 10, "release_date": "2025-10-28", "format": "mlx", "mlx_only": true, "collection": "MiniMax-M2", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3.5-4B-MTP-bf16", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 373, "hf_likes": 1, "release_date": "2026-06-01", "format": "mlx", "mlx_only": true, "collection": "Qwen 3.x MTP", "description": "MLX MTP drafter checkpoints for Qwen 3.x speculative decoding with mlx-vlm.", "_discovered": true }, { "name": "mlx-community/Llama-4-Scout-17B-16E-Instruct-6bit", "provider": "mlx-community", "parameter_count": "17B", "parameters_raw": 17000000000, "min_ram_gb": 15.7, "recommended_ram_gb": 19.2, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 358, "hf_likes": 5, "release_date": "2025-05-03", "format": "mlx", "mlx_only": true, "collection": "Llama 4", "description": "", "_discovered": true }, { "name": "mlx-community/medgemma-1.5-4b-it-8bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 357, "hf_likes": 2, "release_date": "2026-01-14", "format": "mlx", "mlx_only": true, "collection": "MedGemma-1.5", "description": "MedGemma-1.5 models in MLX format. See original repo: https://huggingface.co/google/medgemma-1.5-4b-it", "_discovered": true }, { "name": "mlx-community/Codestral-22B-v0.1-8bit", "provider": "mlx-community", "parameter_count": "22B", "parameters_raw": 22000000000, "min_ram_gb": 26.3, "recommended_ram_gb": 31.7, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 351, "hf_likes": 8, "release_date": "2024-05-29", "format": "mlx", "mlx_only": true, "collection": "Mistral (Mamba) Codestral", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-E2B-it-qat-8bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 351, "hf_likes": 0, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Gemma 4 QAT", "description": "", "_discovered": true }, { "name": "mlx-community/MiMo-V2.5-ASR-MLX-bf16", "provider": "mlx-community", "parameter_count": "7.62262B", "parameters_raw": 7622619136, "min_ram_gb": 18.5, "recommended_ram_gb": 22.6, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "automatic-speech-recognition", "architecture": "", "hf_downloads": 351, "hf_likes": 0, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "MiMo-V2.5-ASR", "description": "by Xiaomi, converted to MLX", "_discovered": true }, { "name": "mlx-community/MiniMax-M2.7-6bit", "provider": "mlx-community", "parameter_count": "228.69B", "parameters_raw": 228689748992, "min_ram_gb": 198.2, "recommended_ram_gb": 233.5, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 346, "hf_likes": 1, "release_date": "2026-04-12", "format": "mlx", "mlx_only": true, "collection": "MiniMax-M2.7", "description": "", "_discovered": true }, { "name": "mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-4bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 18.2, "recommended_ram_gb": 22.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 327, "hf_likes": 3, "release_date": "2025-06-19", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Qwen3", "description": "Abliterated, and further fine-tuned to be the most uncensored models available. Now in MLX", "_discovered": true }, { "name": "mlx-community/Mixtral-8x7B-Instruct-v0.1", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 320, "hf_likes": 23, "release_date": "2024-05-07", "format": "mlx", "mlx_only": true, "collection": "Mistral", "description": "", "_discovered": true }, { "name": "mlx-community/Nemotron-Cascade-2-30B-A3B-6bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 26.9, "recommended_ram_gb": 32.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 319, "hf_likes": 6, "release_date": "2026-03-20", "format": "mlx", "mlx_only": true, "collection": "Nemotron-Cascade 2", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3.5-9B-MTP-bf16", "provider": "mlx-community", "parameter_count": "9B", "parameters_raw": 9000000000, "min_ram_gb": 21.7, "recommended_ram_gb": 26.3, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 317, "hf_likes": 0, "release_date": "2026-06-01", "format": "mlx", "mlx_only": true, "collection": "Qwen 3.x MTP", "description": "MLX MTP drafter checkpoints for Qwen 3.x speculative decoding with mlx-vlm.", "_discovered": true }, { "name": "mlx-community/gemma-3-12b-it-4bit-DWQ", "provider": "mlx-community", "parameter_count": "12B", "parameters_raw": 12000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 316, "hf_likes": 2, "release_date": "2025-05-18", "format": "mlx", "mlx_only": true, "collection": "Gemma 3 DWQ", "description": "Gemma 3 distilled weight quantized (DWQ) models", "_discovered": true }, { "name": "mlx-community/SmolVLM-Instruct-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 314, "hf_likes": 5, "release_date": "2024-11-29", "format": "mlx", "mlx_only": true, "collection": "Idefics 3 + SmolVLM", "description": "", "_discovered": true }, { "name": "mlx-community/Laguna-XS.2-4bit", "provider": "mlx-community", "parameter_count": "33.4426B", "parameters_raw": 33442607104, "min_ram_gb": 20.2, "recommended_ram_gb": 24.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 311, "hf_likes": 4, "release_date": "2026-04-28", "format": "mlx", "mlx_only": true, "collection": "Poolside Laguna-XS.2", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen2.5-VL-72B-Instruct-3bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 32.0, "recommended_ram_gb": 38.5, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 298, "hf_likes": 5, "release_date": "2025-02-25", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-VL", "description": "", "_discovered": true }, { "name": "mlx-community/MiMo-V2.5-ASR-MLX-4bit", "provider": "mlx-community", "parameter_count": "1.25344B", "parameters_raw": 1253440000, "min_ram_gb": 1.7, "recommended_ram_gb": 2.8, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "automatic-speech-recognition", "architecture": "", "hf_downloads": 296, "hf_likes": 0, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "MiMo-V2.5-ASR", "description": "by Xiaomi, converted to MLX", "_discovered": true }, { "name": "mlx-community/Lance-3B-8bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 290, "hf_likes": 3, "release_date": "2026-05-26", "format": "mlx", "mlx_only": true, "collection": "Lance MLX", "description": "Feature-complete MLX port of ByteDance Lance: t2i, image_edit, x2t_image, t2v, video_edit, x2t_video.", "_discovered": true }, { "name": "mlx-community/granite-4.0-h-1b-6bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.9, "recommended_ram_gb": 3.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 283, "hf_likes": 0, "release_date": "2025-10-28", "format": "mlx", "mlx_only": true, "collection": "Granite 4.0 Nano Language Models", "description": "", "_discovered": true }, { "name": "mlx-community/LFM2-VL-450M-8bit", "provider": "mlx-community", "parameter_count": "450M", "parameters_raw": 450000000, "min_ram_gb": 1.5, "recommended_ram_gb": 2.6, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 282, "hf_likes": 11, "release_date": "2025-08-16", "format": "mlx", "mlx_only": true, "collection": "LFM2-VL", "description": "", "_discovered": true }, { "name": "mlx-community/VibeThinker-3B", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 279, "hf_likes": 0, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "VibeThinker-3B", "description": "MLX conversions of VibeThinker-3B for Apple Silicon Chips. ", "_discovered": true }, { "name": "mlx-community/Llama-4-Scout-17B-16E-Instruct-8bit", "provider": "mlx-community", "parameter_count": "17B", "parameters_raw": 17000000000, "min_ram_gb": 20.5, "recommended_ram_gb": 25.0, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 277, "hf_likes": 4, "release_date": "2025-05-03", "format": "mlx", "mlx_only": true, "collection": "Llama 4", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-4b-it-bf16", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 277, "hf_likes": 1, "release_date": "2025-03-20", "format": "mlx", "mlx_only": true, "collection": "Gemma 3", "description": "A collection of lightweight, state-of-the-art open models built from the same research and technology that powers the Gemini 2.0 models", "_discovered": true }, { "name": "mlx-community/MiniMax-M2.7-5bit", "provider": "mlx-community", "parameter_count": "228.69B", "parameters_raw": 228689748992, "min_ram_gb": 165.4, "recommended_ram_gb": 195.0, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 274, "hf_likes": 2, "release_date": "2026-04-12", "format": "mlx", "mlx_only": true, "collection": "MiniMax-M2.7", "description": "", "_discovered": true }, { "name": "mlx-community/Dia-1.6B", "provider": "mlx-community", "parameter_count": "1.6B", "parameters_raw": 1600000000, "min_ram_gb": 1.9, "recommended_ram_gb": 3.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 265, "hf_likes": 24, "release_date": "2025-04-23", "format": "mlx", "mlx_only": true, "collection": "NariLabs Dia-1.5B", "description": "", "_discovered": true }, { "name": "mlx-community/Llama-3.3-70B-Instruct-6bit", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 61.4, "recommended_ram_gb": 72.9, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 265, "hf_likes": 5, "release_date": "2024-12-06", "format": "mlx", "mlx_only": true, "collection": "Llama 3.3", "description": "", "_discovered": true }, { "name": "mlx-community/Nemotron-Cascade-2-30B-A3B-8bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 35.5, "recommended_ram_gb": 42.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 262, "hf_likes": 8, "release_date": "2026-03-20", "format": "mlx", "mlx_only": true, "collection": "Nemotron-Cascade 2", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3.5-397B-A17B-8bit", "provider": "mlx-community", "parameter_count": "397B", "parameters_raw": 397000000000, "min_ram_gb": 457.5, "recommended_ram_gb": 538.0, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 257, "hf_likes": 5, "release_date": "2026-02-20", "format": "mlx", "mlx_only": true, "collection": "Qwen-3.5", "description": "", "_discovered": true }, { "name": "mlx-community/Apertus-8B-Instruct-2509-bf16", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 19.4, "recommended_ram_gb": 23.6, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 251, "hf_likes": 5, "release_date": "2025-09-03", "format": "mlx", "mlx_only": true, "collection": "Apertus", "description": "SwissAI's Apertus models that support 1k languages", "_discovered": true }, { "name": "mlx-community/GLM-4.6-5bit", "provider": "mlx-community", "parameter_count": "352.798B", "parameters_raw": 352797829024, "min_ram_gb": 254.6, "recommended_ram_gb": 299.7, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 250, "hf_likes": 3, "release_date": "2025-09-30", "format": "mlx", "mlx_only": true, "collection": "GLM-4.6", "description": "", "_discovered": true }, { "name": "mlx-community/DeepSeek-R1-Distill-Qwen-7B-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 248, "hf_likes": 8, "release_date": "2025-02-26", "format": "mlx", "mlx_only": true, "collection": "DeepSeek-R1-Distill", "description": "", "_discovered": true }, { "name": "mlx-community/Hy-MT2-7B-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "translation", "architecture": "", "hf_downloads": 248, "hf_likes": 1, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "Hy-MT2", "description": "MLX conversions of Tencent Hy-MT2 1.8B and 7B.", "_discovered": true }, { "name": "mlx-community/parakeet-ctc-0.6b", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "automatic-speech-recognition", "architecture": "", "hf_downloads": 244, "hf_likes": 2, "release_date": "2025-05-10", "format": "mlx", "mlx_only": true, "collection": "Parakeet", "description": "Nvidia's ASR models, now in MLX!", "_discovered": true }, { "name": "mlx-community/Hy-MT2-1.8B-8bit", "provider": "mlx-community", "parameter_count": "1.8B", "parameters_raw": 1800000000, "min_ram_gb": 3.1, "recommended_ram_gb": 4.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "translation", "architecture": "", "hf_downloads": 243, "hf_likes": 1, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "Hy-MT2", "description": "MLX conversions of Tencent Hy-MT2 1.8B and 7B.", "_discovered": true }, { "name": "mlx-community/gemma-4-E2B-it-qat-5bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.4, "recommended_ram_gb": 3.7, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 243, "hf_likes": 0, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Gemma 4 QAT", "description": "", "_discovered": true }, { "name": "mlx-community/Kimi-VL-A3B-Thinking-8bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 235, "hf_likes": 4, "release_date": "2026-01-27", "format": "mlx", "mlx_only": true, "collection": "Kimi-VL Thinking", "description": "", "_discovered": true }, { "name": "mlx-community/functiongemma-270m-it-4bit", "provider": "mlx-community", "parameter_count": "270M", "parameters_raw": 270000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 235, "hf_likes": 3, "release_date": "2025-12-18", "format": "mlx", "mlx_only": true, "collection": "FunctionGemma", "description": "by Google Deepmind", "_discovered": true }, { "name": "mlx-community/Qwen3.5-4B-MTP-5bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.9, "recommended_ram_gb": 5.4, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 230, "hf_likes": 1, "release_date": "2026-06-01", "format": "mlx", "mlx_only": true, "collection": "Qwen 3.x MTP", "description": "MLX MTP drafter checkpoints for Qwen 3.x speculative decoding with mlx-vlm.", "_discovered": true }, { "name": "mlx-community/medgemma-1.5-4b-it-4bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 227, "hf_likes": 3, "release_date": "2026-01-14", "format": "mlx", "mlx_only": true, "collection": "MedGemma-1.5", "description": "MedGemma-1.5 models in MLX format. See original repo: https://huggingface.co/google/medgemma-1.5-4b-it", "_discovered": true }, { "name": "mlx-community/Meta-Llama-3.1-70B-Instruct-bf16", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 162.0, "recommended_ram_gb": 191.0, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 224, "hf_likes": 3, "release_date": "2024-10-06", "format": "mlx", "mlx_only": true, "collection": "Llama 3.1", "description": "", "_discovered": true }, { "name": "mlx-community/SoulX-Singer-fp32", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-audio", "architecture": "", "hf_downloads": 222, "hf_likes": 1, "release_date": "2026-05-31", "format": "mlx", "mlx_only": true, "collection": "SoulX-Singer MLX", "description": "Apple MLX safetensors checkpoints for Soul-AILab SoulX-Singer and SoulX-Singer-SVC.", "_discovered": true }, { "name": "mlx-community/chatterbox-turbo-4bit", "provider": "mlx-community", "parameter_count": "134.507M", "parameters_raw": 134507202, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 220, "hf_likes": 7, "release_date": "2025-12-17", "format": "mlx", "mlx_only": true, "collection": "Chatterbox TTS", "description": "Chatterbox and Chatterbox Turbo By ResembleAI", "_discovered": true }, { "name": "mlx-community/Llama-3.2-11B-Vision-Instruct-abliterated-4-bit", "provider": "mlx-community", "parameter_count": "11B", "parameters_raw": 11000000000, "min_ram_gb": 7.3, "recommended_ram_gb": 9.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 217, "hf_likes": 1, "release_date": "2024-12-16", "format": "mlx", "mlx_only": true, "collection": "Llama 3.2", "description": "Meta goes small with Llama3.2, both text only 1B and 3B, and the 11B Vision models.", "_discovered": true }, { "name": "mlx-community/Qwen2.5-VL-72B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 83.8, "recommended_ram_gb": 99.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 213, "hf_likes": 2, "release_date": "2025-02-25", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-VL", "description": "", "_discovered": true }, { "name": "mlx-community/codegemma-7b-it-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 211, "hf_likes": 6, "release_date": "2024-04-09", "format": "mlx", "mlx_only": true, "collection": "Code Gemma", "description": "Google’s Code-Gemma", "_discovered": true }, { "name": "mlx-community/Qwen3-VL-4B-Instruct-3bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 211, "hf_likes": 4, "release_date": "2025-10-14", "format": "mlx", "mlx_only": true, "collection": "Qwen3-VL", "description": "", "_discovered": true }, { "name": "mlx-community/Agents-A1-8bit", "provider": "mlx-community", "parameter_count": "10.1957B", "parameters_raw": 10195701616, "min_ram_gb": 12.7, "recommended_ram_gb": 15.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 192, "hf_likes": 4, "release_date": "2026-07-01", "format": "mlx", "mlx_only": true, "collection": "Agents-A1", "description": "MLX versions of InternScience/Agents-A1", "_discovered": true }, { "name": "mlx-community/MiniCPM-V-4.6-8bit", "provider": "mlx-community", "parameter_count": "1.07885B", "parameters_raw": 1078850800, "min_ram_gb": 2.2, "recommended_ram_gb": 3.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 192, "hf_likes": 0, "release_date": "2026-05-13", "format": "mlx", "mlx_only": true, "collection": "MiniCPM-V 4.6", "description": "MLX variants of MiniCPM-V 4.6, 1.3B parameters (SigLIP2 400M vision encoder + Qwen3.5-0.8B LLM), repo: https://huggingface.co/openbmb/MiniCPM-V-4.6", "_discovered": true }, { "name": "mlx-community/Meta-Llama-3.1-70B-bf16", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 162.0, "recommended_ram_gb": 191.0, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 187, "hf_likes": 4, "release_date": "2024-07-23", "format": "mlx", "mlx_only": true, "collection": "Llama 3.1", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-2-9b-8bit", "provider": "mlx-community", "parameter_count": "9B", "parameters_raw": 9000000000, "min_ram_gb": 11.3, "recommended_ram_gb": 14.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 186, "hf_likes": 9, "release_date": "2024-06-27", "format": "mlx", "mlx_only": true, "collection": "Google Gemma2", "description": "", "_discovered": true }, { "name": "mlx-community/DiffuCoder-7B-cpGRPO-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 180, "hf_likes": 10, "release_date": "2025-07-04", "format": "mlx", "mlx_only": true, "collection": "DiffuCoder-7B", "description": "Apple's text based diffusion model", "_discovered": true }, { "name": "mlx-community/Agents-A1-5bit", "provider": "mlx-community", "parameter_count": "6.94835B", "parameters_raw": 6948351856, "min_ram_gb": 6.0, "recommended_ram_gb": 7.9, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 178, "hf_likes": 3, "release_date": "2026-07-01", "format": "mlx", "mlx_only": true, "collection": "Agents-A1", "description": "MLX versions of InternScience/Agents-A1", "_discovered": true }, { "name": "mlx-community/Agents-A1-bf16", "provider": "mlx-community", "parameter_count": "35.1072B", "parameters_raw": 35107181936, "min_ram_gb": 81.7, "recommended_ram_gb": 96.8, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 177, "hf_likes": 2, "release_date": "2026-07-01", "format": "mlx", "mlx_only": true, "collection": "Agents-A1", "description": "MLX versions of InternScience/Agents-A1", "_discovered": true }, { "name": "mlx-community/GLM-4.5-4bit", "provider": "mlx-community", "parameter_count": "352.798B", "parameters_raw": 352797829024, "min_ram_gb": 203.9, "recommended_ram_gb": 240.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 176, "hf_likes": 16, "release_date": "2025-07-28", "format": "mlx", "mlx_only": true, "collection": "GLM 4.5", "description": "", "_discovered": true }, { "name": "mlx-community/Molmo-7B-D-0924-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 176, "hf_likes": 2, "release_date": "2024-12-27", "format": "mlx", "mlx_only": true, "collection": "Molmo", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-E2B-it-qat-6bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 173, "hf_likes": 0, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Gemma 4 QAT", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-27b-it-qat-8bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 32.0, "recommended_ram_gb": 38.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 170, "hf_likes": 9, "release_date": "2025-04-19", "format": "mlx", "mlx_only": true, "collection": "Gemma 3 QAT", "description": "Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory.", "_discovered": true }, { "name": "mlx-community/Qwen3-ASR-0.6B-6bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.5, "recommended_ram_gb": 2.6, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 166, "hf_likes": 0, "release_date": "2026-01-29", "format": "mlx", "mlx_only": true, "collection": "Qwen3-ASR", "description": "This collection contains Qwen3-ASR & Qwen3-ForceAligner", "_discovered": true }, { "name": "mlx-community/QwQ-32B-4bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 19.4, "recommended_ram_gb": 23.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 165, "hf_likes": 38, "release_date": "2025-03-05", "format": "mlx", "mlx_only": true, "collection": "Qwen QwQ", "description": "", "_discovered": true }, { "name": "mlx-community/bitnet-b1.58-2B-4T", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 162, "hf_likes": 6, "release_date": "2025-06-10", "format": "mlx", "mlx_only": true, "collection": "BitNet 1.58", "description": "This collection houses BitNet-1.58, Falcon3-1.58 and Falcon-E quants.", "_discovered": true }, { "name": "mlx-community/OmniVoice-4bit", "provider": "mlx-community", "parameter_count": "315.892M", "parameters_raw": 315891912, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 162, "hf_likes": 2, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "OmniVoice", "description": "by k2-fsa, converted to MLX", "_discovered": true }, { "name": "mlx-community/OmniVoice", "provider": "mlx-community", "parameter_count": "612.577M", "parameters_raw": 612577288, "min_ram_gb": 1.4, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 162, "hf_likes": 1, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "OmniVoice", "description": "by k2-fsa, converted to MLX", "_discovered": true }, { "name": "mlx-community/Florence-2-base-ft-4bit", "provider": "mlx-community", "parameter_count": "48.7871M", "parameters_raw": 48787088, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 162, "hf_likes": 1, "release_date": "2024-11-21", "format": "mlx", "mlx_only": true, "collection": "Florence-2", "description": "", "_discovered": true }, { "name": "mlx-community/Llama-3.2-11B-Vision-Instruct-abliterated-8-bit", "provider": "mlx-community", "parameter_count": "11B", "parameters_raw": 11000000000, "min_ram_gb": 13.6, "recommended_ram_gb": 16.9, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 162, "hf_likes": 1, "release_date": "2024-12-16", "format": "mlx", "mlx_only": true, "collection": "Llama 3.2", "description": "Meta goes small with Llama3.2, both text only 1B and 3B, and the 11B Vision models.", "_discovered": true }, { "name": "mlx-community/olmOCR-7B-0725-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 161, "hf_likes": 5, "release_date": "2025-07-25", "format": "mlx", "mlx_only": true, "collection": "olmOCR-0725", "description": "", "_discovered": true }, { "name": "mlx-community/Agents-A1-6bit", "provider": "mlx-community", "parameter_count": "8.0308B", "parameters_raw": 8030801776, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 160, "hf_likes": 2, "release_date": "2026-07-01", "format": "mlx", "mlx_only": true, "collection": "Agents-A1", "description": "MLX versions of InternScience/Agents-A1", "_discovered": true }, { "name": "mlx-community/Step-3.5-Flash-8bit", "provider": "mlx-community", "parameter_count": "196.956B", "parameters_raw": 196956118272, "min_ram_gb": 227.5, "recommended_ram_gb": 267.9, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 155, "hf_likes": 2, "release_date": "2026-02-04", "format": "mlx", "mlx_only": true, "collection": "Step 3.5 Flash", "description": "By StepFun", "_discovered": true }, { "name": "mlx-community/Hermes-2-Pro-Mistral-7B-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 154, "hf_likes": 5, "release_date": "2024-03-14", "format": "mlx", "mlx_only": true, "collection": "Mistral", "description": "", "_discovered": true }, { "name": "mlx-community/Devstral-Small-2505-3bit", "provider": "mlx-community", "parameter_count": "2.94691B", "parameters_raw": 2946913280, "min_ram_gb": 2.3, "recommended_ram_gb": 3.5, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 153, "hf_likes": 1, "release_date": "2025-05-21", "format": "mlx", "mlx_only": true, "collection": "Devstral Small 2505", "description": "", "_discovered": true }, { "name": "mlx-community/Bernini-R-int4", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-video", "architecture": "", "hf_downloads": 152, "hf_likes": 6, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "Bernini-R MLX", "description": "MLX port of ByteDance Bernini-R: Wan2.2-A14B video renderer/editor with SA-3D RoPE (t2v/r2v/v2v/rv2v). Renderer-only, UMT5 conditioning.", "_discovered": true }, { "name": "mlx-community/Meta-Llama-3-70B-4bit", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 41.2, "recommended_ram_gb": 49.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 149, "hf_likes": 9, "release_date": "2024-04-20", "format": "mlx", "mlx_only": true, "collection": "Llama 3", "description": "", "_discovered": true }, { "name": "mlx-community/VibeVoice-Realtime-0.5B-4bit", "provider": "mlx-community", "parameter_count": "500M", "parameters_raw": 500000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 149, "hf_likes": 7, "release_date": "2025-12-15", "format": "mlx", "mlx_only": true, "collection": "VibeVoice", "description": "", "_discovered": true }, { "name": "mlx-community/Granite-4.0-H-Tiny-4bit-DWQ", "provider": "mlx-community", "parameter_count": "1.08542B", "parameters_raw": 1085424192, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 146, "hf_likes": 5, "release_date": "2025-10-03", "format": "mlx", "mlx_only": true, "collection": "Granite-4.0 Family", "description": "", "_discovered": true }, { "name": "mlx-community/Real-ESRGAN-x4plus", "provider": "mlx-community", "parameter_count": "16.698M", "parameters_raw": 16697987, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 141, "hf_likes": 3, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "Real-ESRGAN (MLX)", "description": "Apple MLX fp16 ports of Real-ESRGAN super-resolution (RRDBNet + SRVGGNetCompact), 5 variants, BSD-3.", "_discovered": true }, { "name": "mlx-community/Yi-1.5-34B-Chat-8bit", "provider": "mlx-community", "parameter_count": "34B", "parameters_raw": 34000000000, "min_ram_gb": 40.1, "recommended_ram_gb": 47.9, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 140, "hf_likes": 3, "release_date": "2024-05-13", "format": "mlx", "mlx_only": true, "collection": "Yi-1.5", "description": "", "_discovered": true }, { "name": "mlx-community/DeepSeek-R1-0528-Qwen3-8B-4bit-DWQ", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 139, "hf_likes": 8, "release_date": "2025-05-29", "format": "mlx", "mlx_only": true, "collection": "DeepSeek R1 0528", "description": "", "_discovered": true }, { "name": "mlx-community/Llama-OuteTTS-1.0-1B-4bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 139, "hf_likes": 1, "release_date": "2025-05-19", "format": "mlx", "mlx_only": true, "collection": "OuteTTS-1.0", "description": "", "_discovered": true }, { "name": "mlx-community/olmOCR-2-7B-1025-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 138, "hf_likes": 2, "release_date": "2025-10-22", "format": "mlx", "mlx_only": true, "collection": "olmOCR 2", "description": "", "_discovered": true }, { "name": "mlx-community/Nanonets-OCR-s-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 135, "hf_likes": 2, "release_date": "2025-06-18", "format": "mlx", "mlx_only": true, "collection": "Nanonets OCR", "description": "This collection houses Nanonets-OCR-s", "_discovered": true }, { "name": "mlx-community/granite-4.0-tiny-preview-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 135, "hf_likes": 0, "release_date": "2025-09-10", "format": "mlx", "mlx_only": true, "collection": "Granite-4.0 Family", "description": "", "_discovered": true }, { "name": "mlx-community/Mistral-7B-v0.2-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 130, "hf_likes": 8, "release_date": "2024-03-25", "format": "mlx", "mlx_only": true, "collection": "Mistral", "description": "", "_discovered": true }, { "name": "mlx-community/Real-ESRGAN-x4plus-anime-6B", "provider": "mlx-community", "parameter_count": "6B", "parameters_raw": 6000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 130, "hf_likes": 3, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "Real-ESRGAN (MLX)", "description": "Apple MLX fp16 ports of Real-ESRGAN super-resolution (RRDBNet + SRVGGNetCompact), 5 variants, BSD-3.", "_discovered": true }, { "name": "mlx-community/Qwen3.5-397B-A17B-6bit", "provider": "mlx-community", "parameter_count": "397B", "parameters_raw": 397000000000, "min_ram_gb": 343.4, "recommended_ram_gb": 404.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 130, "hf_likes": 2, "release_date": "2026-02-19", "format": "mlx", "mlx_only": true, "collection": "Qwen-3.5", "description": "", "_discovered": true }, { "name": "mlx-community/OmniVoice-8bit", "provider": "mlx-community", "parameter_count": "622.148M", "parameters_raw": 622147784, "min_ram_gb": 1.7, "recommended_ram_gb": 2.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 130, "hf_likes": 1, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "OmniVoice", "description": "by k2-fsa, converted to MLX", "_discovered": true }, { "name": "mlx-community/gemma-2-27b-it-8bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 32.0, "recommended_ram_gb": 38.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 127, "hf_likes": 10, "release_date": "2024-11-06", "format": "mlx", "mlx_only": true, "collection": "Google Gemma2", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-4-31b-6bit", "provider": "mlx-community", "parameter_count": "31B", "parameters_raw": 31000000000, "min_ram_gb": 27.7, "recommended_ram_gb": 33.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 125, "hf_likes": 2, "release_date": "2026-04-02", "format": "mlx", "mlx_only": true, "collection": "Gemma 4", "description": "", "_discovered": true }, { "name": "mlx-community/mel-roformer-kim-vocal-2-mlx", "provider": "mlx-community", "parameter_count": "228.203M", "parameters_raw": 228203172, "min_ram_gb": 1.1, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 123, "hf_likes": 6, "release_date": "2026-05-01", "format": "mlx", "mlx_only": true, "collection": "Mel-Band-RoFormer (MLX)", "description": "MLX-format Mel-Band-RoFormer vocal source separation models (MIT-licensed, parity-tested vs PyTorch reference)", "_discovered": true }, { "name": "mlx-community/Apriel-1.5-15b-Thinker-4bit", "provider": "mlx-community", "parameter_count": "15B", "parameters_raw": 15000000000, "min_ram_gb": 9.6, "recommended_ram_gb": 12.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 121, "hf_likes": 2, "release_date": "2025-10-03", "format": "mlx", "mlx_only": true, "collection": "ServiceNow-Apriel", "description": "Apriel-1.5-15b-Thinker is a multimodal reasoning model in ServiceNow’s Apriel SLM series which achieves competitive performance against models 10 time", "_discovered": true }, { "name": "mlx-community/gemma-3-27b-it-qat-bf16", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 63.1, "recommended_ram_gb": 74.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 120, "hf_likes": 6, "release_date": "2025-04-18", "format": "mlx", "mlx_only": true, "collection": "Gemma 3 QAT", "description": "Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory.", "_discovered": true }, { "name": "mlx-community/GLM-4-32B-0414-4bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 19.4, "recommended_ram_gb": 23.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 119, "hf_likes": 4, "release_date": "2025-04-21", "format": "mlx", "mlx_only": true, "collection": "GLM4", "description": "The GLM-4 and Z1 series are powerful open-source language models excelling in reasoning, code, and complex tasks.", "_discovered": true }, { "name": "mlx-community/Qwen3-Coder-30B-A3B-Instruct-8bit-DWQ-lr9e8", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 35.5, "recommended_ram_gb": 42.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 119, "hf_likes": 1, "release_date": "2025-08-01", "format": "mlx", "mlx_only": true, "collection": "Qwen3-Coder-MoE", "description": "💻 Significant Performance: among open models on Agentic Coding, Agentic Browser-Use, and other foundational coding tasks, achieving ~Claude Sonnet.", "_discovered": true }, { "name": "mlx-community/Lens-3.8B-bf16", "provider": "mlx-community", "parameter_count": "3.8B", "parameters_raw": 3800000000, "min_ram_gb": 9.7, "recommended_ram_gb": 12.3, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-image", "architecture": "", "hf_downloads": 115, "hf_likes": 3, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Lens 3.8B (MLX)", "description": "Apple MLX conversions of microsoft/Lens — 3.8B text-to-image DiT (GPT-OSS features + FLUX.2 VAE) for Apple Silicon. bf16 + int4/int8.", "_discovered": true }, { "name": "mlx-community/granite-4.0-h-micro-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 115, "hf_likes": 2, "release_date": "2025-10-02", "format": "mlx", "mlx_only": true, "collection": "Granite-4.0 Family", "description": "", "_discovered": true }, { "name": "mlx-community/whisper-tiny-mlx-q4", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 115, "hf_likes": 2, "release_date": "2024-03-09", "format": "mlx", "mlx_only": true, "collection": "Whisper", "description": "OpenAI Whisper speech recognition models in MLX format", "_discovered": true }, { "name": "mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-8bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 35.5, "recommended_ram_gb": 42.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 115, "hf_likes": 1, "release_date": "2025-06-19", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Qwen3", "description": "Abliterated, and further fine-tuned to be the most uncensored models available. Now in MLX", "_discovered": true }, { "name": "mlx-community/LFM2-350M-4bit", "provider": "mlx-community", "parameter_count": "350M", "parameters_raw": 350000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 111, "hf_likes": 6, "release_date": "2025-07-11", "format": "mlx", "mlx_only": true, "collection": "LFM2.x", "description": "", "_discovered": true }, { "name": "mlx-community/plamo-2-1b", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 111, "hf_likes": 4, "release_date": "2025-03-15", "format": "mlx", "mlx_only": true, "collection": "PLaMo", "description": "", "_discovered": true }, { "name": "mlx-community/Mamba-Codestral-7B-v0.1-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 111, "hf_likes": 2, "release_date": "2025-01-21", "format": "mlx", "mlx_only": true, "collection": "Mamba2", "description": "", "_discovered": true }, { "name": "mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 111, "hf_likes": 1, "release_date": "2025-02-16", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Qwen2.5", "description": "The best uncensored models", "_discovered": true }, { "name": "mlx-community/SeedVR2-3B-mlx-int8", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 109, "hf_likes": 2, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "SeedVR2 (MLX-Swift)", "description": "SeedVR2-3B (ByteDance, ICLR 2026) one-step diffusion super-resolution, MLX-Swift weights for on-device Apple Silicon. fp16 + int8.", "_discovered": true }, { "name": "mlx-community/granite-4.1-30b-bf16", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 70.0, "recommended_ram_gb": 83.0, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 108, "hf_likes": 1, "release_date": "2026-04-30", "format": "mlx", "mlx_only": true, "collection": "Granite 4.1", "description": "By IBM", "_discovered": true }, { "name": "mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 107, "hf_likes": 1, "release_date": "2025-02-16", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Qwen2.5", "description": "The best uncensored models", "_discovered": true }, { "name": "mlx-community/gemma-3n-E4B-it-lm-bf16", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 106, "hf_likes": 5, "release_date": "2025-06-29", "format": "mlx", "mlx_only": true, "collection": "Gemma 3n - Text Only (LM)", "description": "Google's Gemma 3n converted to MLX using mlx-lm", "_discovered": true }, { "name": "mlx-community/EfRLFN-x4", "provider": "mlx-community", "parameter_count": "503.894K", "parameters_raw": 503894, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 105, "hf_likes": 7, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "EfRLFN MLX", "description": "MLX port of EfRLFN (ICLR 2026): realtime x2/x4 image super-resolution on Apple Silicon.", "_discovered": true }, { "name": "mlx-community/encodec-32khz-float32", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 105, "hf_likes": 0, "release_date": "2024-09-18", "format": "mlx", "mlx_only": true, "collection": "EnCodec", "description": "EnCodec models in MLX", "_discovered": true }, { "name": "mlx-community/Ministral-8B-Instruct-2410-8bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 104, "hf_likes": 2, "release_date": "2024-10-17", "format": "mlx", "mlx_only": true, "collection": "Ministral", "description": "", "_discovered": true }, { "name": "mlx-community/PaddleOCR-VL-8bit", "provider": "mlx-community", "parameter_count": "351.728M", "parameters_raw": 351727700, "min_ram_gb": 1.4, "recommended_ram_gb": 2.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 101, "hf_likes": 3, "release_date": "2026-01-19", "format": "mlx", "mlx_only": true, "collection": "PaddleOCR-VL", "description": "", "_discovered": true }, { "name": "mlx-community/Mistral-Small-24B-Instruct-2501-3bit", "provider": "mlx-community", "parameter_count": "24B", "parameters_raw": 24000000000, "min_ram_gb": 11.3, "recommended_ram_gb": 14.2, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 101, "hf_likes": 2, "release_date": "2025-01-30", "format": "mlx", "mlx_only": true, "collection": "Mistral Small", "description": "", "_discovered": true }, { "name": "mlx-community/Nanonets-OCR2-3B-8bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 99, "hf_likes": 1, "release_date": "2025-10-14", "format": "mlx", "mlx_only": true, "collection": "Nanonets OCR2", "description": "This collection houses Nanonets-OCR2 models", "_discovered": true }, { "name": "mlx-community/SeedVR2-3B-mlx", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 98, "hf_likes": 2, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "SeedVR2 (MLX-Swift)", "description": "SeedVR2-3B (ByteDance, ICLR 2026) one-step diffusion super-resolution, MLX-Swift weights for on-device Apple Silicon. fp16 + int8.", "_discovered": true }, { "name": "mlx-community/Mamba-Codestral-7B-v0.1-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 98, "hf_likes": 1, "release_date": "2025-01-21", "format": "mlx", "mlx_only": true, "collection": "Mamba2", "description": "", "_discovered": true }, { "name": "mlx-community/Mamba-Codestral-7B-v0.1", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 97, "hf_likes": 2, "release_date": "2025-01-21", "format": "mlx", "mlx_only": true, "collection": "Mamba2", "description": "", "_discovered": true }, { "name": "mlx-community/Kokoro-82M-6bit", "provider": "mlx-community", "parameter_count": "82M", "parameters_raw": 82000000, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 96, "hf_likes": 2, "release_date": "2026-01-05", "format": "mlx", "mlx_only": true, "collection": "Kokoro TTS", "description": "Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers amazing quality.", "_discovered": true }, { "name": "mlx-community/deepseek-vl2-small-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 96, "hf_likes": 0, "release_date": "2024-12-22", "format": "mlx", "mlx_only": true, "collection": "DeepSeek-VL2", "description": "", "_discovered": true }, { "name": "mlx-community/Devstral-Small-2505-8bit", "provider": "mlx-community", "parameter_count": "6.63004B", "parameters_raw": 6630036480, "min_ram_gb": 8.6, "recommended_ram_gb": 11.0, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 92, "hf_likes": 2, "release_date": "2025-05-21", "format": "mlx", "mlx_only": true, "collection": "Devstral Small 2505", "description": "", "_discovered": true }, { "name": "mlx-community/Laguna-XS.2-8bit", "provider": "mlx-community", "parameter_count": "33.4426B", "parameters_raw": 33442607104, "min_ram_gb": 39.5, "recommended_ram_gb": 47.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 91, "hf_likes": 1, "release_date": "2026-04-28", "format": "mlx", "mlx_only": true, "collection": "Poolside Laguna-XS.2", "description": "", "_discovered": true }, { "name": "mlx-community/Apertus-8B-Instruct-2509-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 91, "hf_likes": 1, "release_date": "2025-09-03", "format": "mlx", "mlx_only": true, "collection": "Apertus", "description": "SwissAI's Apertus models that support 1k languages", "_discovered": true }, { "name": "mlx-community/Ministral-8B-Instruct-2410-bf16", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 19.4, "recommended_ram_gb": 23.6, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 89, "hf_likes": 2, "release_date": "2024-10-17", "format": "mlx", "mlx_only": true, "collection": "Ministral", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen2.5-VL-72B-Instruct-6bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 63.1, "recommended_ram_gb": 74.9, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 89, "hf_likes": 1, "release_date": "2025-02-25", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-VL", "description": "", "_discovered": true }, { "name": "mlx-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-6bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 26.9, "recommended_ram_gb": 32.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 89, "hf_likes": 0, "release_date": "2026-04-29", "format": "mlx", "mlx_only": true, "collection": "Nvidia Nemotron-3-Nano-Omni", "description": "", "_discovered": true }, { "name": "mlx-community/Orchestrator-8B-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 87, "hf_likes": 5, "release_date": "2025-11-29", "format": "mlx", "mlx_only": true, "collection": "Orchestrator 8B", "description": "", "_discovered": true }, { "name": "mlx-community/bitnet-b1.58-2B-4T-4bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 87, "hf_likes": 0, "release_date": "2025-06-10", "format": "mlx", "mlx_only": true, "collection": "BitNet 1.58", "description": "This collection houses BitNet-1.58, Falcon3-1.58 and Falcon-E quants.", "_discovered": true }, { "name": "mlx-community/OmniVoice-fp32", "provider": "mlx-community", "parameter_count": "612.577M", "parameters_raw": 612577288, "min_ram_gb": 1.4, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 86, "hf_likes": 1, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "OmniVoice", "description": "by k2-fsa, converted to MLX", "_discovered": true }, { "name": "mlx-community/gemma-4-31b-5bit", "provider": "mlx-community", "parameter_count": "31B", "parameters_raw": 31000000000, "min_ram_gb": 23.3, "recommended_ram_gb": 28.2, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 85, "hf_likes": 2, "release_date": "2026-04-02", "format": "mlx", "mlx_only": true, "collection": "Gemma 4", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16-mlx-6Bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 63.1, "recommended_ram_gb": 74.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 84, "hf_likes": 5, "release_date": "2026-05-13", "format": "mlx", "mlx_only": true, "collection": "UNCENSORED Qwen 3.6 27B", "description": "", "_discovered": true }, { "name": "mlx-community/MiniMax-M2-3bit", "provider": "mlx-community", "parameter_count": "228.69B", "parameters_raw": 228689748992, "min_ram_gb": 99.6, "recommended_ram_gb": 117.8, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 84, "hf_likes": 3, "release_date": "2025-10-28", "format": "mlx", "mlx_only": true, "collection": "MiniMax-M2", "description": "", "_discovered": true }, { "name": "mlx-community/OpenELM-1_1B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 83, "hf_likes": 4, "release_date": "2024-04-24", "format": "mlx", "mlx_only": true, "collection": "OpenELM", "description": "A family of Open-source Efficient Language Models from Apple.", "_discovered": true }, { "name": "mlx-community/GLM-Z1-32B-0414-4bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 19.4, "recommended_ram_gb": 23.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 82, "hf_likes": 2, "release_date": "2025-04-19", "format": "mlx", "mlx_only": true, "collection": "GLM4", "description": "The GLM-4 and Z1 series are powerful open-source language models excelling in reasoning, code, and complex tasks.", "_discovered": true }, { "name": "mlx-community/Qwen2.5-7B-Instruct-1M-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 81, "hf_likes": 4, "release_date": "2025-01-26", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-1M", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-TTS-12Hz-0.6B-Base-6bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.5, "recommended_ram_gb": 2.6, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 81, "hf_likes": 2, "release_date": "2026-01-25", "format": "mlx", "mlx_only": true, "collection": "Qwen3-TTS", "description": "", "_discovered": true }, { "name": "mlx-community/sam-audio-large", "provider": "mlx-community", "parameter_count": "3.04081B", "parameters_raw": 3040807045, "min_ram_gb": 2.7, "recommended_ram_gb": 4.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 80, "hf_likes": 7, "release_date": "2025-12-24", "format": "mlx", "mlx_only": true, "collection": "Sam Audio", "description": "By Facebook ", "_discovered": true }, { "name": "mlx-community/functiongemma-270m-it-bf16", "provider": "mlx-community", "parameter_count": "270M", "parameters_raw": 270000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 80, "hf_likes": 7, "release_date": "2025-12-18", "format": "mlx", "mlx_only": true, "collection": "FunctionGemma", "description": "by Google Deepmind", "_discovered": true }, { "name": "mlx-community/GLM-4.5-Air-mxfp4", "provider": "mlx-community", "parameter_count": "106.852B", "parameters_raw": 106852251264, "min_ram_gb": 62.4, "recommended_ram_gb": 74.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 80, "hf_likes": 2, "release_date": "2025-09-26", "format": "mlx", "mlx_only": true, "collection": "GLM-4.5-Air", "description": "", "_discovered": true }, { "name": "mlx-community/Meta-Llama-3-8B-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 78, "hf_likes": 8, "release_date": "2024-04-20", "format": "mlx", "mlx_only": true, "collection": "Llama 3", "description": "", "_discovered": true }, { "name": "mlx-community/Bernini-R-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-video", "architecture": "", "hf_downloads": 76, "hf_likes": 2, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "Bernini-R MLX", "description": "MLX port of ByteDance Bernini-R: Wan2.2-A14B video renderer/editor with SA-3D RoPE (t2v/r2v/v2v/rv2v). Renderer-only, UMT5 conditioning.", "_discovered": true }, { "name": "mlx-community/LFM2-8B-A1B-3bit-MLX", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 76, "hf_likes": 2, "release_date": "2025-10-08", "format": "mlx", "mlx_only": true, "collection": "💧LFM2-8B-A1B-MoE", "description": "Best in Class MoE, better than Qwen3. Optimised for Smaller devices sub 16 GB (M1/2/3/4) Apple Silicon.", "_discovered": true }, { "name": "mlx-community/OLMoE-1B-7B-0125-Instruct-4bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 76, "hf_likes": 2, "release_date": "2025-03-04", "format": "mlx", "mlx_only": true, "collection": "OLMoE", "description": "", "_discovered": true }, { "name": "mlx-community/Step-3.5-Flash-6bit", "provider": "mlx-community", "parameter_count": "196.956B", "parameters_raw": 196956118272, "min_ram_gb": 170.9, "recommended_ram_gb": 201.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 75, "hf_likes": 1, "release_date": "2026-02-04", "format": "mlx", "mlx_only": true, "collection": "Step 3.5 Flash", "description": "By StepFun", "_discovered": true }, { "name": "mlx-community/Nanonets-OCR2-3B-bf16", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 75, "hf_likes": 0, "release_date": "2025-10-14", "format": "mlx", "mlx_only": true, "collection": "Nanonets OCR2", "description": "This collection houses Nanonets-OCR2 models", "_discovered": true }, { "name": "mlx-community/QwQ-32B-8bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 37.8, "recommended_ram_gb": 45.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 73, "hf_likes": 14, "release_date": "2025-03-05", "format": "mlx", "mlx_only": true, "collection": "Qwen QwQ", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-4B-Instruct-2507-gabliterated-8bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 73, "hf_likes": 1, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "Gabliterated v1", "description": "The next version of Abliteration", "_discovered": true }, { "name": "mlx-community/codegemma-7b-it-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 72, "hf_likes": 2, "release_date": "2024-04-09", "format": "mlx", "mlx_only": true, "collection": "Code Gemma", "description": "Google’s Code-Gemma", "_discovered": true }, { "name": "mlx-community/Lens-3.8B-8bit", "provider": "mlx-community", "parameter_count": "3.8B", "parameters_raw": 3800000000, "min_ram_gb": 5.4, "recommended_ram_gb": 7.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-image", "architecture": "", "hf_downloads": 72, "hf_likes": 1, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Lens 3.8B (MLX)", "description": "Apple MLX conversions of microsoft/Lens — 3.8B text-to-image DiT (GPT-OSS features + FLUX.2 VAE) for Apple Silicon. bf16 + int4/int8.", "_discovered": true }, { "name": "mlx-community/mamba-790m-hf-f16", "provider": "mlx-community", "parameter_count": "790M", "parameters_raw": 790000000, "min_ram_gb": 1.5, "recommended_ram_gb": 2.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 72, "hf_likes": 0, "release_date": "2024-09-21", "format": "mlx", "mlx_only": true, "collection": "Mamba", "description": "Mamba is a new LLM architecture that integrates the Structured State Space sequence model to manage lengthy data sequences.", "_discovered": true }, { "name": "mlx-community/GLM-4.5-Air-3bit-DWQ-v2", "provider": "mlx-community", "parameter_count": "106.852B", "parameters_raw": 106852251264, "min_ram_gb": 47.1, "recommended_ram_gb": 56.1, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 69, "hf_likes": 4, "release_date": "2025-08-13", "format": "mlx", "mlx_only": true, "collection": "GLM-4.5-Air", "description": "", "_discovered": true }, { "name": "mlx-community/Mellum-4b-base", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 69, "hf_likes": 2, "release_date": "2025-06-28", "format": "mlx", "mlx_only": true, "collection": "JetBrains Mellum", "description": "Series of code models by JetBrains", "_discovered": true }, { "name": "mlx-community/Lens-3.8B-4bit", "provider": "mlx-community", "parameter_count": "3.8B", "parameters_raw": 3800000000, "min_ram_gb": 3.2, "recommended_ram_gb": 4.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-image", "architecture": "", "hf_downloads": 69, "hf_likes": 1, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Lens 3.8B (MLX)", "description": "Apple MLX conversions of microsoft/Lens — 3.8B text-to-image DiT (GPT-OSS features + FLUX.2 VAE) for Apple Silicon. bf16 + int4/int8.", "_discovered": true }, { "name": "mlx-community/Devstral-Small-2505-6bit", "provider": "mlx-community", "parameter_count": "5.15679B", "parameters_raw": 5156787200, "min_ram_gb": 5.4, "recommended_ram_gb": 7.2, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 69, "hf_likes": 1, "release_date": "2025-05-21", "format": "mlx", "mlx_only": true, "collection": "Devstral Small 2505", "description": "", "_discovered": true }, { "name": "mlx-community/DiffuCoder-7B-cpGRPO-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 68, "hf_likes": 5, "release_date": "2025-07-04", "format": "mlx", "mlx_only": true, "collection": "DiffuCoder-7B", "description": "Apple's text based diffusion model", "_discovered": true }, { "name": "mlx-community/PaddleOCR-VL-4bit", "provider": "mlx-community", "parameter_count": "255.402M", "parameters_raw": 255401796, "min_ram_gb": 1.1, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 68, "hf_likes": 2, "release_date": "2026-01-19", "format": "mlx", "mlx_only": true, "collection": "PaddleOCR-VL", "description": "", "_discovered": true }, { "name": "mlx-community/Josiefied-Olmo-3-7B-Instruct-abliterated-v1-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 68, "hf_likes": 0, "release_date": "2025-11-25", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Models", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-27b-it-4bit-DWQ", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 16.5, "recommended_ram_gb": 20.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 67, "hf_likes": 3, "release_date": "2025-05-14", "format": "mlx", "mlx_only": true, "collection": "Gemma 3 DWQ", "description": "Gemma 3 distilled weight quantized (DWQ) models", "_discovered": true }, { "name": "mlx-community/DeepSeek-R1-0528-Qwen3-8B-8bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 67, "hf_likes": 2, "release_date": "2025-05-29", "format": "mlx", "mlx_only": true, "collection": "DeepSeek R1 0528", "description": "", "_discovered": true }, { "name": "mlx-community/LFM2-VL-450M-4bit", "provider": "mlx-community", "parameter_count": "450M", "parameters_raw": 450000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 66, "hf_likes": 1, "release_date": "2025-08-16", "format": "mlx", "mlx_only": true, "collection": "LFM2-VL", "description": "", "_discovered": true }, { "name": "mlx-community/medgemma-4b-it-8bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 65, "hf_likes": 3, "release_date": "2025-06-09", "format": "mlx", "mlx_only": true, "collection": "MedGemma", "description": "Collection of Gemma 3 variants for performance on medical text and image comprehension to accelerate building healthcare-based AI applications.", "_discovered": true }, { "name": "mlx-community/IQuest-Coder-V1-40B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "40B", "parameters_raw": 40000000000, "min_ram_gb": 47.0, "recommended_ram_gb": 56.0, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 64, "hf_likes": 6, "release_date": "2026-01-01", "format": "mlx", "mlx_only": true, "collection": "IQuest-Coder", "description": "By IQuestLab", "_discovered": true }, { "name": "mlx-community/Real-ESRGAN-x2plus", "provider": "mlx-community", "parameter_count": "16.7032M", "parameters_raw": 16703171, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 64, "hf_likes": 1, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "Real-ESRGAN (MLX)", "description": "Apple MLX fp16 ports of Real-ESRGAN super-resolution (RRDBNet + SRVGGNetCompact), 5 variants, BSD-3.", "_discovered": true }, { "name": "mlx-community/Qwen2.5-7B-Instruct-1M-3bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 4.0, "recommended_ram_gb": 5.5, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 63, "hf_likes": 0, "release_date": "2025-01-26", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-1M", "description": "", "_discovered": true }, { "name": "mlx-community/mel-roformer-zfturbo-vocals-v1-mlx", "provider": "mlx-community", "parameter_count": "33.6674M", "parameters_raw": 33667396, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 62, "hf_likes": 2, "release_date": "2026-05-01", "format": "mlx", "mlx_only": true, "collection": "Mel-Band-RoFormer (MLX)", "description": "MLX-format Mel-Band-RoFormer vocal source separation models (MIT-licensed, parity-tested vs PyTorch reference)", "_discovered": true }, { "name": "mlx-community/Nanonets-OCR2-3B-4bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 62, "hf_likes": 0, "release_date": "2025-10-14", "format": "mlx", "mlx_only": true, "collection": "Nanonets OCR2", "description": "This collection houses Nanonets-OCR2 models", "_discovered": true }, { "name": "mlx-community/VibeVoice-Realtime-0.5B-8bit", "provider": "mlx-community", "parameter_count": "500M", "parameters_raw": 500000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 61, "hf_likes": 3, "release_date": "2025-12-15", "format": "mlx", "mlx_only": true, "collection": "VibeVoice", "description": "", "_discovered": true }, { "name": "mlx-community/Laguna-XS.2-6bit", "provider": "mlx-community", "parameter_count": "33.4426B", "parameters_raw": 33442607104, "min_ram_gb": 29.8, "recommended_ram_gb": 35.9, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 61, "hf_likes": 0, "release_date": "2026-04-28", "format": "mlx", "mlx_only": true, "collection": "Poolside Laguna-XS.2", "description": "", "_discovered": true }, { "name": "mlx-community/AceReason-Nemotron-7B-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 61, "hf_likes": 0, "release_date": "2025-05-26", "format": "mlx", "mlx_only": true, "collection": "AceReason Nemotron", "description": "", "_discovered": true }, { "name": "mlx-community/Falcon3-Mamba-7B-Instruct-4bits", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 61, "hf_likes": 0, "release_date": "2025-02-14", "format": "mlx", "mlx_only": true, "collection": "Falcon3 Mamba", "description": "", "_discovered": true }, { "name": "mlx-community/Yi-1.5-9B-Chat-4bit", "provider": "mlx-community", "parameter_count": "9B", "parameters_raw": 9000000000, "min_ram_gb": 6.2, "recommended_ram_gb": 8.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 60, "hf_likes": 2, "release_date": "2024-05-13", "format": "mlx", "mlx_only": true, "collection": "Yi-1.5", "description": "", "_discovered": true }, { "name": "mlx-community/SigLIP2-NR-IQA-KonIQ", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-classification", "architecture": "", "hf_downloads": 60, "hf_likes": 1, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "SigLIP2 NR-IQA MLX", "description": "MLX no-reference image-quality head on SigLIP2-SO400M (repro of arXiv:2509.17374).", "_discovered": true }, { "name": "mlx-community/Florence-2-large-ft-bf16", "provider": "mlx-community", "parameter_count": "822.899M", "parameters_raw": 822898688, "min_ram_gb": 2.9, "recommended_ram_gb": 4.2, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 60, "hf_likes": 1, "release_date": "2024-11-21", "format": "mlx", "mlx_only": true, "collection": "Florence-2", "description": "", "_discovered": true }, { "name": "mlx-community/Dia-1.6B-4bit", "provider": "mlx-community", "parameter_count": "1.6B", "parameters_raw": 1600000000, "min_ram_gb": 1.9, "recommended_ram_gb": 3.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 59, "hf_likes": 13, "release_date": "2025-04-24", "format": "mlx", "mlx_only": true, "collection": "NariLabs Dia-1.5B", "description": "", "_discovered": true }, { "name": "mlx-community/Laguna-XS.2-5bit", "provider": "mlx-community", "parameter_count": "33.4426B", "parameters_raw": 33442607104, "min_ram_gb": 25.0, "recommended_ram_gb": 30.2, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 59, "hf_likes": 0, "release_date": "2026-04-28", "format": "mlx", "mlx_only": true, "collection": "Poolside Laguna-XS.2", "description": "", "_discovered": true }, { "name": "mlx-community/Phi-3-mini-128k-instruct-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 58, "hf_likes": 10, "release_date": "2024-07-11", "format": "mlx", "mlx_only": true, "collection": "Phi-3", "description": "", "_discovered": true }, { "name": "mlx-community/Meta-Llama-Guard-2-8B-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 58, "hf_likes": 0, "release_date": "2024-04-19", "format": "mlx", "mlx_only": true, "collection": "Llama 3", "description": "", "_discovered": true }, { "name": "mlx-community/V-JEPA2-vitl-fpc64-256", "provider": "mlx-community", "parameter_count": "325.971M", "parameters_raw": 325971328, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "video-classification", "architecture": "", "hf_downloads": 57, "hf_likes": 1, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "V-JEPA2 (MLX)", "description": "Apple MLX fp16 ports of Meta V-JEPA2 ViT-L — video embeddings, JEPA predictor, SSv2 classifier. MIT.", "_discovered": true }, { "name": "mlx-community/Kimi-VL-A3B-Thinking-6bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 3.6, "recommended_ram_gb": 5.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 57, "hf_likes": 1, "release_date": "2026-01-27", "format": "mlx", "mlx_only": true, "collection": "Kimi-VL Thinking", "description": "", "_discovered": true }, { "name": "mlx-community/SoulX-Singer-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-audio", "architecture": "", "hf_downloads": 57, "hf_likes": 0, "release_date": "2026-05-31", "format": "mlx", "mlx_only": true, "collection": "SoulX-Singer MLX", "description": "Apple MLX safetensors checkpoints for Soul-AILab SoulX-Singer and SoulX-Singer-SVC.", "_discovered": true }, { "name": "mlx-community/Cocktail-Fork-MRX-adapted-loudness", "provider": "mlx-community", "parameter_count": "30.5664M", "parameters_raw": 30566448, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 56, "hf_likes": 1, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Cocktail-Fork MRX (MLX)", "description": "MERL MRX ported to Apple MLX — 3-stem music/speech/sfx soundtrack separation. Numerically exact vs PyTorch. 4 variants.", "_discovered": true }, { "name": "mlx-community/Qwen3-8B-4bit-DWQ-053125", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 55, "hf_likes": 3, "release_date": "2025-06-01", "format": "mlx", "mlx_only": true, "collection": "Qwen3 DWQ Quants", "description": "High-quality 4-bit quants of the Qwen3 model family.", "_discovered": true }, { "name": "mlx-community/sam-audio-large-fp16", "provider": "mlx-community", "parameter_count": "3.04081B", "parameters_raw": 3040807045, "min_ram_gb": 2.7, "recommended_ram_gb": 4.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 54, "hf_likes": 6, "release_date": "2025-12-24", "format": "mlx", "mlx_only": true, "collection": "Sam Audio", "description": "By Facebook ", "_discovered": true }, { "name": "mlx-community/embeddinggemma-300m-5bit", "provider": "mlx-community", "parameter_count": "300M", "parameters_raw": 300000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "embedding", "capabilities": [ "mlx" ], "pipeline_tag": "sentence-similarity", "architecture": "", "hf_downloads": 54, "hf_likes": 0, "release_date": "2025-09-04", "format": "mlx", "mlx_only": true, "collection": "EmbeddingGemma", "description": "", "_discovered": true }, { "name": "mlx-community/Mistral-Nemo-Base-2407-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 53, "hf_likes": 3, "release_date": "2024-07-18", "format": "mlx", "mlx_only": true, "collection": "Mistral NeMo", "description": "", "_discovered": true }, { "name": "mlx-community/OpenELM-1_1B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 53, "hf_likes": 2, "release_date": "2024-04-24", "format": "mlx", "mlx_only": true, "collection": "OpenELM", "description": "A family of Open-source Efficient Language Models from Apple.", "_discovered": true }, { "name": "mlx-community/Qwen2.5-72B-Instruct-bf16", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 166.6, "recommended_ram_gb": 196.4, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 53, "hf_likes": 0, "release_date": "2024-09-19", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5", "description": "The Qwen 2.5 models are a series of AI models trained on 18 trillion tokens, supporting 29 languages and offering advanced features such as instructio", "_discovered": true }, { "name": "mlx-community/Mellum-4b-base-8bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 52, "hf_likes": 1, "release_date": "2025-06-28", "format": "mlx", "mlx_only": true, "collection": "JetBrains Mellum", "description": "Series of code models by JetBrains", "_discovered": true }, { "name": "mlx-community/Cocktail-Fork-MRX", "provider": "mlx-community", "parameter_count": "30.5664M", "parameters_raw": 30566448, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 50, "hf_likes": 1, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Cocktail-Fork MRX (MLX)", "description": "MERL MRX ported to Apple MLX — 3-stem music/speech/sfx soundtrack separation. Numerically exact vs PyTorch. 4 variants.", "_discovered": true }, { "name": "mlx-community/Florence-2-base-ft-8bit", "provider": "mlx-community", "parameter_count": "81.6936M", "parameters_raw": 81693648, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 50, "hf_likes": 1, "release_date": "2024-11-21", "format": "mlx", "mlx_only": true, "collection": "Florence-2", "description": "", "_discovered": true }, { "name": "mlx-community/Mistral-Small-24B-Instruct-2501-8bit", "provider": "mlx-community", "parameter_count": "24B", "parameters_raw": 24000000000, "min_ram_gb": 28.6, "recommended_ram_gb": 34.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 49, "hf_likes": 3, "release_date": "2025-01-30", "format": "mlx", "mlx_only": true, "collection": "Mistral Small", "description": "", "_discovered": true }, { "name": "mlx-community/V-JEPA2-vitl-fpc16-256-ssv2", "provider": "mlx-community", "parameter_count": "353.4M", "parameters_raw": 353399982, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "video-classification", "architecture": "", "hf_downloads": 49, "hf_likes": 1, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "V-JEPA2 (MLX)", "description": "Apple MLX fp16 ports of Meta V-JEPA2 ViT-L — video embeddings, JEPA predictor, SSv2 classifier. MIT.", "_discovered": true }, { "name": "mlx-community/codegemma-2b-4bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 49, "hf_likes": 1, "release_date": "2024-04-09", "format": "mlx", "mlx_only": true, "collection": "Code Gemma", "description": "Google’s Code-Gemma", "_discovered": true }, { "name": "mlx-community/Phi-3-mini-4k-instruct-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 48, "hf_likes": 2, "release_date": "2024-07-04", "format": "mlx", "mlx_only": true, "collection": "Phi-3", "description": "", "_discovered": true }, { "name": "mlx-community/V-JEPA2-AC-vitg", "provider": "mlx-community", "parameter_count": "1.31739B", "parameters_raw": 1317394944, "min_ram_gb": 1.8, "recommended_ram_gb": 2.9, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "video-classification", "architecture": "", "hf_downloads": 48, "hf_likes": 1, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "V-JEPA2 (MLX)", "description": "Apple MLX fp16 ports of Meta V-JEPA2 ViT-L — video embeddings, JEPA predictor, SSv2 classifier. MIT.", "_discovered": true }, { "name": "mlx-community/Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16-mlx-5Bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 63.1, "recommended_ram_gb": 74.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 48, "hf_likes": 0, "release_date": "2026-05-13", "format": "mlx", "mlx_only": true, "collection": "UNCENSORED Qwen 3.6 27B", "description": "", "_discovered": true }, { "name": "mlx-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-5bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 22.6, "recommended_ram_gb": 27.3, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 48, "hf_likes": 0, "release_date": "2026-04-29", "format": "mlx", "mlx_only": true, "collection": "Nvidia Nemotron-3-Nano-Omni", "description": "", "_discovered": true }, { "name": "mlx-community/kitten-tts-nano-0.8", "provider": "mlx-community", "parameter_count": "14.5913M", "parameters_raw": 14591314, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 48, "hf_likes": 0, "release_date": "2026-02-24", "format": "mlx", "mlx_only": true, "collection": "KittenTTS", "description": "All MLX conversions of KittenTTS (nano/micro/mini) across fp32, fp16, bf16, and 4/5/6/8-bit quantizations.", "_discovered": true }, { "name": "mlx-community/SmolVLM-Instruct-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 47, "hf_likes": 9, "release_date": "2024-11-29", "format": "mlx", "mlx_only": true, "collection": "Idefics 3 + SmolVLM", "description": "", "_discovered": true }, { "name": "mlx-community/Real-ESRGAN-animevideov3", "provider": "mlx-community", "parameter_count": "621.424K", "parameters_raw": 621424, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 47, "hf_likes": 1, "release_date": "2026-06-06", "format": "mlx", "mlx_only": true, "collection": "Real-ESRGAN (MLX)", "description": "Apple MLX fp16 ports of Real-ESRGAN super-resolution (RRDBNet + SRVGGNetCompact), 5 variants, BSD-3.", "_discovered": true }, { "name": "mlx-community/MiniCPM-V-4.6-5bit", "provider": "mlx-community", "parameter_count": "1.04995B", "parameters_raw": 1049949424, "min_ram_gb": 1.8, "recommended_ram_gb": 2.9, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 47, "hf_likes": 0, "release_date": "2026-05-13", "format": "mlx", "mlx_only": true, "collection": "MiniCPM-V 4.6", "description": "MLX variants of MiniCPM-V 4.6, 1.3B parameters (SigLIP2 400M vision encoder + Qwen3.5-0.8B LLM), repo: https://huggingface.co/openbmb/MiniCPM-V-4.6", "_discovered": true }, { "name": "mlx-community/Qwen3-ASR-0.6B-5bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.4, "recommended_ram_gb": 2.5, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 47, "hf_likes": 0, "release_date": "2026-01-29", "format": "mlx", "mlx_only": true, "collection": "Qwen3-ASR", "description": "This collection contains Qwen3-ASR & Qwen3-ForceAligner", "_discovered": true }, { "name": "mlx-community/LFM2-8B-A1B-8bit-MLX", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 46, "hf_likes": 7, "release_date": "2025-10-08", "format": "mlx", "mlx_only": true, "collection": "💧LFM2-8B-A1B-MoE", "description": "Best in Class MoE, better than Qwen3. Optimised for Smaller devices sub 16 GB (M1/2/3/4) Apple Silicon.", "_discovered": true }, { "name": "mlx-community/Orchestrator-8B-8bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 46, "hf_likes": 5, "release_date": "2025-11-29", "format": "mlx", "mlx_only": true, "collection": "Orchestrator 8B", "description": "", "_discovered": true }, { "name": "mlx-community/medgemma-1.5-4b-it-6bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 46, "hf_likes": 2, "release_date": "2026-01-14", "format": "mlx", "mlx_only": true, "collection": "MedGemma-1.5", "description": "MedGemma-1.5 models in MLX format. See original repo: https://huggingface.co/google/medgemma-1.5-4b-it", "_discovered": true }, { "name": "mlx-community/Hy3-preview-6bit", "provider": "mlx-community", "parameter_count": "295.034B", "parameters_raw": 295033528320, "min_ram_gb": 255.5, "recommended_ram_gb": 300.7, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 46, "hf_likes": 1, "release_date": "2026-04-27", "format": "mlx", "mlx_only": true, "collection": "Hy3 preview", "description": "By Tencent", "_discovered": true }, { "name": "mlx-community/GLM-4.6V-Flash-5bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 6.0, "recommended_ram_gb": 7.9, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 46, "hf_likes": 0, "release_date": "2025-12-08", "format": "mlx", "mlx_only": true, "collection": "GLM-4.6V", "description": "", "_discovered": true }, { "name": "mlx-community/GLM-4.6V-Flash-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 45, "hf_likes": 2, "release_date": "2025-12-08", "format": "mlx", "mlx_only": true, "collection": "GLM-4.6V", "description": "", "_discovered": true }, { "name": "mlx-community/Hy3-preview-8bit", "provider": "mlx-community", "parameter_count": "295.034B", "parameters_raw": 295033528320, "min_ram_gb": 340.3, "recommended_ram_gb": 400.3, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 44, "hf_likes": 1, "release_date": "2026-04-27", "format": "mlx", "mlx_only": true, "collection": "Hy3 preview", "description": "By Tencent", "_discovered": true }, { "name": "mlx-community/olmOCR-2-7B-1025-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 44, "hf_likes": 0, "release_date": "2025-10-25", "format": "mlx", "mlx_only": true, "collection": "olmOCR 2", "description": "", "_discovered": true }, { "name": "mlx-community/ERNIE-4.5-300B-A47B-PT-4bit", "provider": "mlx-community", "parameter_count": "300B", "parameters_raw": 300000000000, "min_ram_gb": 173.5, "recommended_ram_gb": 204.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 43, "hf_likes": 2, "release_date": "2025-07-04", "format": "mlx", "mlx_only": true, "collection": "ERNIE-4.5", "description": "", "_discovered": true }, { "name": "mlx-community/Apertus-8B-Instruct-2509-8bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 43, "hf_likes": 1, "release_date": "2025-09-03", "format": "mlx", "mlx_only": true, "collection": "Apertus", "description": "SwissAI's Apertus models that support 1k languages", "_discovered": true }, { "name": "mlx-community/Qwen3.5-397B-A17B-5bit", "provider": "mlx-community", "parameter_count": "397B", "parameters_raw": 397000000000, "min_ram_gb": 286.3, "recommended_ram_gb": 337.0, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 43, "hf_likes": 0, "release_date": "2026-02-19", "format": "mlx", "mlx_only": true, "collection": "Qwen-3.5", "description": "", "_discovered": true }, { "name": "mlx-community/bitnet-b1.58-2B-4T-8bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 43, "hf_likes": 0, "release_date": "2025-06-10", "format": "mlx", "mlx_only": true, "collection": "BitNet 1.58", "description": "This collection houses BitNet-1.58, Falcon3-1.58 and Falcon-E quants.", "_discovered": true }, { "name": "mlx-community/granite-4.0-h-1b-4bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 42, "hf_likes": 2, "release_date": "2025-10-28", "format": "mlx", "mlx_only": true, "collection": "Granite 4.0 Nano Language Models", "description": "", "_discovered": true }, { "name": "mlx-community/SmolLM3-3B-8bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 41, "hf_likes": 9, "release_date": "2025-07-08", "format": "mlx", "mlx_only": true, "collection": "SmolLM3", "description": "", "_discovered": true }, { "name": "mlx-community/deepseek-vl2-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 41, "hf_likes": 2, "release_date": "2024-12-22", "format": "mlx", "mlx_only": true, "collection": "DeepSeek-VL2", "description": "", "_discovered": true }, { "name": "mlx-community/lille-130m-instruct-fp16", "provider": "mlx-community", "parameter_count": "130M", "parameters_raw": 130000000, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 41, "hf_likes": 1, "release_date": "2025-09-05", "format": "mlx", "mlx_only": true, "collection": "Lille 130M", "description": "Very Small smart model created for the mobile", "_discovered": true }, { "name": "mlx-community/parakeet-ctc-1.1b", "provider": "mlx-community", "parameter_count": "1.1B", "parameters_raw": 1100000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "automatic-speech-recognition", "architecture": "", "hf_downloads": 41, "hf_likes": 1, "release_date": "2025-05-10", "format": "mlx", "mlx_only": true, "collection": "Parakeet", "description": "Nvidia's ASR models, now in MLX!", "_discovered": true }, { "name": "mlx-community/YOLO26n-OptiQ-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "object-detection", "architecture": "", "hf_downloads": 41, "hf_likes": 0, "release_date": "2026-04-26", "format": "mlx", "mlx_only": true, "collection": "YOLO 26", "description": "", "_discovered": true }, { "name": "mlx-community/SmolVLM-Instruct-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 40, "hf_likes": 5, "release_date": "2024-11-29", "format": "mlx", "mlx_only": true, "collection": "Idefics 3 + SmolVLM", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-4b-pt-4bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 40, "hf_likes": 3, "release_date": "2025-03-18", "format": "mlx", "mlx_only": true, "collection": "Gemma 3", "description": "A collection of lightweight, state-of-the-art open models built from the same research and technology that powers the Gemini 2.0 models", "_discovered": true }, { "name": "mlx-community/Florence-2-base-ft-bf16", "provider": "mlx-community", "parameter_count": "270.906M", "parameters_raw": 270906368, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 40, "hf_likes": 1, "release_date": "2024-11-21", "format": "mlx", "mlx_only": true, "collection": "Florence-2", "description": "", "_discovered": true }, { "name": "mlx-community/Laguna-XS-2.1-bf16", "provider": "mlx-community", "parameter_count": "33.4426B", "parameters_raw": 33442617088, "min_ram_gb": 77.9, "recommended_ram_gb": 92.3, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 40, "hf_likes": 0, "release_date": "2026-07-03", "format": "mlx", "mlx_only": true, "collection": "Laguna-XS-2.1", "description": "MLX versions of Laguna-XS-2.1", "_discovered": true }, { "name": "mlx-community/Qwen3-0.6B-6bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.5, "recommended_ram_gb": 2.6, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 39, "hf_likes": 1, "release_date": "2025-04-28", "format": "mlx", "mlx_only": true, "collection": "Qwen3", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3n-E2B-it-lm-bf16", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 39, "hf_likes": 0, "release_date": "2025-06-29", "format": "mlx", "mlx_only": true, "collection": "Gemma 3n - Text Only (LM)", "description": "Google's Gemma 3n converted to MLX using mlx-lm", "_discovered": true }, { "name": "mlx-community/Ling-2.6-flash-mlx-5bit", "provider": "mlx-community", "parameter_count": "104.187B", "parameters_raw": 104186907648, "min_ram_gb": 75.9, "recommended_ram_gb": 89.9, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 38, "hf_likes": 1, "release_date": "2026-04-30", "format": "mlx", "mlx_only": true, "collection": "inclusionAI LING 2.6", "description": "By inclusionAI", "_discovered": true }, { "name": "mlx-community/Qwen3-Next-80B-A3B-Thinking-6bit", "provider": "mlx-community", "parameter_count": "80B", "parameters_raw": 80000000000, "min_ram_gb": 70.0, "recommended_ram_gb": 83.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 38, "hf_likes": 1, "release_date": "2025-09-13", "format": "mlx", "mlx_only": true, "collection": "Qwen3 Next", "description": "Alibaba's first hybrid model, designed to cut resources and speed things up.", "_discovered": true }, { "name": "mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 38, "hf_likes": 1, "release_date": "2025-02-16", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Qwen2.5", "description": "The best uncensored models", "_discovered": true }, { "name": "mlx-community/olmOCR-2-7B-1025-5bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 6.0, "recommended_ram_gb": 7.9, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 38, "hf_likes": 0, "release_date": "2025-10-25", "format": "mlx", "mlx_only": true, "collection": "olmOCR 2", "description": "", "_discovered": true }, { "name": "mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-6bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 26.9, "recommended_ram_gb": 32.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 38, "hf_likes": 0, "release_date": "2025-06-19", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Qwen3", "description": "Abliterated, and further fine-tuned to be the most uncensored models available. Now in MLX", "_discovered": true }, { "name": "mlx-community/paligemma2-10b-ft-docci-448-bf16", "provider": "mlx-community", "parameter_count": "10B", "parameters_raw": 10000000000, "min_ram_gb": 24.0, "recommended_ram_gb": 29.0, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 36, "hf_likes": 3, "release_date": "2024-12-16", "format": "mlx", "mlx_only": true, "collection": "Paligemma 2", "description": "", "_discovered": true }, { "name": "mlx-community/IQuest-Coder-V1-40B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "40B", "parameters_raw": 40000000000, "min_ram_gb": 24.0, "recommended_ram_gb": 29.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 36, "hf_likes": 2, "release_date": "2026-01-01", "format": "mlx", "mlx_only": true, "collection": "IQuest-Coder", "description": "By IQuestLab", "_discovered": true }, { "name": "mlx-community/VoxCPM1.5-4bit", "provider": "mlx-community", "parameter_count": "211.425M", "parameters_raw": 211424577, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 36, "hf_likes": 1, "release_date": "2025-12-16", "format": "mlx", "mlx_only": true, "collection": "VoxCPM", "description": "", "_discovered": true }, { "name": "mlx-community/olmOCR-7B-0225-preview-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 35, "hf_likes": 4, "release_date": "2025-03-03", "format": "mlx", "mlx_only": true, "collection": "olmOCR", "description": "", "_discovered": true }, { "name": "mlx-community/LLaDA2.0-mini-4bit", "provider": "mlx-community", "parameter_count": "16.2556B", "parameters_raw": 16255643392, "min_ram_gb": 10.3, "recommended_ram_gb": 13.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 35, "hf_likes": 3, "release_date": "2025-11-25", "format": "mlx", "mlx_only": true, "collection": "LLaDA 2.0", "description": "", "_discovered": true }, { "name": "mlx-community/GLM-4.6V-Flash-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 35, "hf_likes": 1, "release_date": "2025-12-08", "format": "mlx", "mlx_only": true, "collection": "GLM-4.6V", "description": "", "_discovered": true }, { "name": "mlx-community/medgemma-4b-it-bf16", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 10.2, "recommended_ram_gb": 12.8, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 35, "hf_likes": 1, "release_date": "2025-06-09", "format": "mlx", "mlx_only": true, "collection": "MedGemma", "description": "Collection of Gemma 3 variants for performance on medical text and image comprehension to accelerate building healthcare-based AI applications.", "_discovered": true }, { "name": "mlx-community/parakeet-rnnt-1.1b", "provider": "mlx-community", "parameter_count": "1.1B", "parameters_raw": 1100000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "automatic-speech-recognition", "architecture": "", "hf_downloads": 35, "hf_likes": 1, "release_date": "2025-05-10", "format": "mlx", "mlx_only": true, "collection": "Parakeet", "description": "Nvidia's ASR models, now in MLX!", "_discovered": true }, { "name": "mlx-community/YOLO26s-OptiQ-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "object-detection", "architecture": "", "hf_downloads": 35, "hf_likes": 0, "release_date": "2026-04-26", "format": "mlx", "mlx_only": true, "collection": "YOLO 26", "description": "", "_discovered": true }, { "name": "mlx-community/lille-130m-instruct-bf16", "provider": "mlx-community", "parameter_count": "130M", "parameters_raw": 130000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 35, "hf_likes": 0, "release_date": "2025-09-05", "format": "mlx", "mlx_only": true, "collection": "Lille 130M", "description": "Very Small smart model created for the mobile", "_discovered": true }, { "name": "mlx-community/Solar-Open-100B-4bit", "provider": "mlx-community", "parameter_count": "100B", "parameters_raw": 100000000000, "min_ram_gb": 58.5, "recommended_ram_gb": 69.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 34, "hf_likes": 0, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "Solar Open", "description": "A 102B-parameter Mixture-of-Experts model by Upstage", "_discovered": true }, { "name": "mlx-community/parakeet-rnnt-0.6b", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "automatic-speech-recognition", "architecture": "", "hf_downloads": 34, "hf_likes": 0, "release_date": "2025-05-10", "format": "mlx", "mlx_only": true, "collection": "Parakeet", "description": "Nvidia's ASR models, now in MLX!", "_discovered": true }, { "name": "mlx-community/sam-audio-small", "provider": "mlx-community", "parameter_count": "602.312M", "parameters_raw": 602312324, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 33, "hf_likes": 4, "release_date": "2025-12-23", "format": "mlx", "mlx_only": true, "collection": "Sam Audio", "description": "By Facebook ", "_discovered": true }, { "name": "mlx-community/Ling-2.6-flash-mlx-4bit", "provider": "mlx-community", "parameter_count": "104.187B", "parameters_raw": 104186907648, "min_ram_gb": 60.9, "recommended_ram_gb": 72.3, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 33, "hf_likes": 2, "release_date": "2026-04-29", "format": "mlx", "mlx_only": true, "collection": "inclusionAI LING 2.6", "description": "By inclusionAI", "_discovered": true }, { "name": "mlx-community/codegemma-7b-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 33, "hf_likes": 1, "release_date": "2024-04-09", "format": "mlx", "mlx_only": true, "collection": "Code Gemma", "description": "Google’s Code-Gemma", "_discovered": true }, { "name": "mlx-community/SoulX-Singer-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-audio", "architecture": "", "hf_downloads": 33, "hf_likes": 0, "release_date": "2026-05-31", "format": "mlx", "mlx_only": true, "collection": "SoulX-Singer MLX", "description": "Apple MLX safetensors checkpoints for Soul-AILab SoulX-Singer and SoulX-Singer-SVC.", "_discovered": true }, { "name": "mlx-community/P1-VL-30B-A3B-8bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 35.5, "recommended_ram_gb": 42.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 33, "hf_likes": 0, "release_date": "2026-05-18", "format": "mlx", "mlx_only": true, "collection": "PRIME-RL P1-VL-30B-A3B", "description": "Bridging visual perception and scientific reasoning in physics olympiads", "_discovered": true }, { "name": "mlx-community/Solar-Open-100B-8bit", "provider": "mlx-community", "parameter_count": "100B", "parameters_raw": 100000000000, "min_ram_gb": 116.0, "recommended_ram_gb": 137.0, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 33, "hf_likes": 0, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "Solar Open", "description": "A 102B-parameter Mixture-of-Experts model by Upstage", "_discovered": true }, { "name": "mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 33, "hf_likes": 0, "release_date": "2025-02-16", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Qwen2.5", "description": "The best uncensored models", "_discovered": true }, { "name": "mlx-community/Qwen1.5-1.8B-Chat-4bit", "provider": "mlx-community", "parameter_count": "1.8B", "parameters_raw": 1800000000, "min_ram_gb": 2.0, "recommended_ram_gb": 3.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 32, "hf_likes": 2, "release_date": "2024-02-18", "format": "mlx", "mlx_only": true, "collection": "Qwen1.5", "description": "Qwen1.5 is the improved version of Qwen, the large language model series developed by Alibaba Cloud.", "_discovered": true }, { "name": "mlx-community/Falcon3-Mamba-7B-Instruct-8bits", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 32, "hf_likes": 0, "release_date": "2025-02-14", "format": "mlx", "mlx_only": true, "collection": "Falcon3 Mamba", "description": "", "_discovered": true }, { "name": "mlx-community/VoxCPM1.5", "provider": "mlx-community", "parameter_count": "887.786M", "parameters_raw": 887786241, "min_ram_gb": 1.5, "recommended_ram_gb": 2.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 31, "hf_likes": 4, "release_date": "2025-12-10", "format": "mlx", "mlx_only": true, "collection": "VoxCPM", "description": "", "_discovered": true }, { "name": "mlx-community/MiniMax-M2-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 31, "hf_likes": 2, "release_date": "2025-10-28", "format": "mlx", "mlx_only": true, "collection": "MiniMax-M2", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen2.5-7B-Instruct-1M-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 31, "hf_likes": 2, "release_date": "2025-01-26", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5-1M", "description": "", "_discovered": true }, { "name": "mlx-community/Cocktail-Fork-MRX-paper", "provider": "mlx-community", "parameter_count": "30.5664M", "parameters_raw": 30566448, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 31, "hf_likes": 1, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Cocktail-Fork MRX (MLX)", "description": "MERL MRX ported to Apple MLX — 3-stem music/speech/sfx soundtrack separation. Numerically exact vs PyTorch. 4 variants.", "_discovered": true }, { "name": "mlx-community/Lens-Turbo-3.8B-bf16", "provider": "mlx-community", "parameter_count": "3.8B", "parameters_raw": 3800000000, "min_ram_gb": 9.7, "recommended_ram_gb": 12.3, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-image", "architecture": "", "hf_downloads": 31, "hf_likes": 0, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Lens 3.8B (MLX)", "description": "Apple MLX conversions of microsoft/Lens — 3.8B text-to-image DiT (GPT-OSS features + FLUX.2 VAE) for Apple Silicon. bf16 + int4/int8.", "_discovered": true }, { "name": "mlx-community/YOLO26m-OptiQ-6bit", "provider": "mlx-community", "parameter_count": "26M", "parameters_raw": 26000000, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "object-detection", "architecture": "", "hf_downloads": 31, "hf_likes": 0, "release_date": "2026-04-26", "format": "mlx", "mlx_only": true, "collection": "YOLO 26", "description": "", "_discovered": true }, { "name": "mlx-community/granite-4.0-h-small-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 31, "hf_likes": 0, "release_date": "2025-10-02", "format": "mlx", "mlx_only": true, "collection": "Granite-4.0 Family", "description": "", "_discovered": true }, { "name": "mlx-community/Nanonets-OCR-s-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 31, "hf_likes": 0, "release_date": "2025-07-25", "format": "mlx", "mlx_only": true, "collection": "Nanonets OCR", "description": "This collection houses Nanonets-OCR-s", "_discovered": true }, { "name": "mlx-community/EXAONE-3.5-2.4B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "2.4B", "parameters_raw": 2400000000, "min_ram_gb": 2.4, "recommended_ram_gb": 3.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 30, "hf_likes": 3, "release_date": "2024-12-09", "format": "mlx", "mlx_only": true, "collection": "EXAONE-3.5", "description": "EXAONE 3.5, a collection of instruction-tuned bilingual generative models ranging from 2.4B to 32B parameters, developed by LG AI.", "_discovered": true }, { "name": "mlx-community/MiniMax-M2-5bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 6.0, "recommended_ram_gb": 7.9, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 30, "hf_likes": 2, "release_date": "2025-10-29", "format": "mlx", "mlx_only": true, "collection": "MiniMax-M2", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-Next-80B-A3B-Thinking-5bit", "provider": "mlx-community", "parameter_count": "80B", "parameters_raw": 80000000000, "min_ram_gb": 58.5, "recommended_ram_gb": 69.5, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 30, "hf_likes": 1, "release_date": "2025-09-13", "format": "mlx", "mlx_only": true, "collection": "Qwen3 Next", "description": "Alibaba's first hybrid model, designed to cut resources and speed things up.", "_discovered": true }, { "name": "mlx-community/SU-01-5bit", "provider": "mlx-community", "parameter_count": "30.5321B", "parameters_raw": 30532122624, "min_ram_gb": 22.9, "recommended_ram_gb": 27.8, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 30, "hf_likes": 0, "release_date": "2026-05-16", "format": "mlx", "mlx_only": true, "collection": "Simplified Reasoning SU-01", "description": "Rigorous mathematical and scientific olympiad problem solving", "_discovered": true }, { "name": "mlx-community/bitnet-b1.58-2B-4T-6bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 30, "hf_likes": 0, "release_date": "2025-06-10", "format": "mlx", "mlx_only": true, "collection": "BitNet 1.58", "description": "This collection houses BitNet-1.58, Falcon3-1.58 and Falcon-E quants.", "_discovered": true }, { "name": "mlx-community/Dia-1.6B-6bit", "provider": "mlx-community", "parameter_count": "1.6B", "parameters_raw": 1600000000, "min_ram_gb": 2.4, "recommended_ram_gb": 3.6, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 29, "hf_likes": 10, "release_date": "2025-04-24", "format": "mlx", "mlx_only": true, "collection": "NariLabs Dia-1.5B", "description": "", "_discovered": true }, { "name": "mlx-community/LFM2-1.2B-8bit", "provider": "mlx-community", "parameter_count": "1.2B", "parameters_raw": 1200000000, "min_ram_gb": 2.4, "recommended_ram_gb": 3.6, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 29, "hf_likes": 2, "release_date": "2025-07-12", "format": "mlx", "mlx_only": true, "collection": "LFM2.x", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-2-27b-8bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 32.0, "recommended_ram_gb": 38.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 29, "hf_likes": 2, "release_date": "2024-06-27", "format": "mlx", "mlx_only": true, "collection": "Google Gemma2", "description": "", "_discovered": true }, { "name": "mlx-community/Cocktail-Fork-MRX-adapted-eq", "provider": "mlx-community", "parameter_count": "30.5664M", "parameters_raw": 30566448, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 29, "hf_likes": 1, "release_date": "2026-06-05", "format": "mlx", "mlx_only": true, "collection": "Cocktail-Fork MRX (MLX)", "description": "MERL MRX ported to Apple MLX — 3-stem music/speech/sfx soundtrack separation. Numerically exact vs PyTorch. 4 variants.", "_discovered": true }, { "name": "mlx-community/Nemotron-Cascade-2-30B-A3B-5bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 22.6, "recommended_ram_gb": 27.3, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 29, "hf_likes": 1, "release_date": "2026-03-20", "format": "mlx", "mlx_only": true, "collection": "Nemotron-Cascade 2", "description": "", "_discovered": true }, { "name": "mlx-community/Mistral-Nemo-Base-2407-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 29, "hf_likes": 1, "release_date": "2024-07-18", "format": "mlx", "mlx_only": true, "collection": "Mistral NeMo", "description": "", "_discovered": true }, { "name": "mlx-community/Falcon3-Mamba-7B-Instruct", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 29, "hf_likes": 0, "release_date": "2025-02-14", "format": "mlx", "mlx_only": true, "collection": "Falcon3 Mamba", "description": "", "_discovered": true }, { "name": "mlx-community/deepseek-vl2-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 27, "hf_likes": 1, "release_date": "2024-12-22", "format": "mlx", "mlx_only": true, "collection": "DeepSeek-VL2", "description": "", "_discovered": true }, { "name": "mlx-community/Laguna-XS-2.1-5bit", "provider": "mlx-community", "parameter_count": "6.27256B", "parameters_raw": 6272558848, "min_ram_gb": 5.5, "recommended_ram_gb": 7.3, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 27, "hf_likes": 0, "release_date": "2026-07-03", "format": "mlx", "mlx_only": true, "collection": "Laguna-XS-2.1", "description": "MLX versions of Laguna-XS-2.1", "_discovered": true }, { "name": "mlx-community/LLaDA2.0-flash-4bit", "provider": "mlx-community", "parameter_count": "102.89B", "parameters_raw": 102889705216, "min_ram_gb": 60.2, "recommended_ram_gb": 71.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 26, "hf_likes": 3, "release_date": "2025-11-25", "format": "mlx", "mlx_only": true, "collection": "LLaDA 2.0", "description": "", "_discovered": true }, { "name": "mlx-community/medgemma-4b-it-6bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 26, "hf_likes": 1, "release_date": "2025-06-09", "format": "mlx", "mlx_only": true, "collection": "MedGemma", "description": "Collection of Gemma 3 variants for performance on medical text and image comprehension to accelerate building healthcare-based AI applications.", "_discovered": true }, { "name": "mlx-community/K-EXAONE-236B-A23B-8bit", "provider": "mlx-community", "parameter_count": "236B", "parameters_raw": 236000000000, "min_ram_gb": 272.4, "recommended_ram_gb": 320.6, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 26, "hf_likes": 0, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "K-EXAONE", "description": " A large-scale multilingual language model by LG AI Research", "_discovered": true }, { "name": "mlx-community/plamo-2-1b-bf16", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 25, "hf_likes": 2, "release_date": "2025-03-15", "format": "mlx", "mlx_only": true, "collection": "PLaMo", "description": "", "_discovered": true }, { "name": "mlx-community/Llama-OuteTTS-1.0-1B-8bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 25, "hf_likes": 1, "release_date": "2025-05-19", "format": "mlx", "mlx_only": true, "collection": "OuteTTS-1.0", "description": "", "_discovered": true }, { "name": "mlx-community/SongGeneration-v2-medium-8bit", "provider": "mlx-community", "parameter_count": "788.837M", "parameters_raw": 788837088, "min_ram_gb": 1.9, "recommended_ram_gb": 3.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-audio", "architecture": "", "hf_downloads": 25, "hf_likes": 0, "release_date": "2026-05-31", "format": "mlx", "mlx_only": true, "collection": "SongGeneration v2 MLX", "description": "Apple MLX checkpoints for Tencent SongGeneration v2 medium and large audiolm token generation.", "_discovered": true }, { "name": "mlx-community/Solar-Open-100B-6bit", "provider": "mlx-community", "parameter_count": "100B", "parameters_raw": 100000000000, "min_ram_gb": 87.2, "recommended_ram_gb": 103.2, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 25, "hf_likes": 0, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "Solar Open", "description": "A 102B-parameter Mixture-of-Experts model by Upstage", "_discovered": true }, { "name": "mlx-community/deepseek-vl2-small-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 25, "hf_likes": 0, "release_date": "2024-12-22", "format": "mlx", "mlx_only": true, "collection": "DeepSeek-VL2", "description": "", "_discovered": true }, { "name": "mlx-community/Lumimaid-8B-v0.1", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 25, "hf_likes": 0, "release_date": "2024-10-13", "format": "mlx", "mlx_only": true, "collection": "Lumimaid", "description": "A collection of Neversleep's RP focused Lumimaid LLMs.", "_discovered": true }, { "name": "mlx-community/Qwen3-4B-4bit-DWQ-053125", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 24, "hf_likes": 2, "release_date": "2025-06-01", "format": "mlx", "mlx_only": true, "collection": "Qwen3 DWQ Quants", "description": "High-quality 4-bit quants of the Qwen3 model family.", "_discovered": true }, { "name": "mlx-community/NAFNet-GoPro-width64", "provider": "mlx-community", "parameter_count": "67.8888M", "parameters_raw": 67888835, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 24, "hf_likes": 1, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "NAFNet MLX", "description": "MLX port of NAFNet (Simple Baselines for Image Restoration): on-device deblur/denoise on Apple Silicon.", "_discovered": true }, { "name": "mlx-community/Apertus-8B-Instruct-2509-6bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 24, "hf_likes": 1, "release_date": "2025-09-03", "format": "mlx", "mlx_only": true, "collection": "Apertus", "description": "SwissAI's Apertus models that support 1k languages", "_discovered": true }, { "name": "mlx-community/DiffuCoder-7B-cpGRPO-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 24, "hf_likes": 1, "release_date": "2025-07-04", "format": "mlx", "mlx_only": true, "collection": "DiffuCoder-7B", "description": "Apple's text based diffusion model", "_discovered": true }, { "name": "mlx-community/Qwen1.5-14B-Chat-4bit", "provider": "mlx-community", "parameter_count": "14B", "parameters_raw": 14000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 24, "hf_likes": 1, "release_date": "2024-03-08", "format": "mlx", "mlx_only": true, "collection": "Qwen1.5", "description": "Qwen1.5 is the improved version of Qwen, the large language model series developed by Alibaba Cloud.", "_discovered": true }, { "name": "mlx-community/Laguna-XS-2.1-6bit", "provider": "mlx-community", "parameter_count": "7.317B", "parameters_raw": 7316995840, "min_ram_gb": 7.3, "recommended_ram_gb": 9.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 24, "hf_likes": 0, "release_date": "2026-07-03", "format": "mlx", "mlx_only": true, "collection": "Laguna-XS-2.1", "description": "MLX versions of Laguna-XS-2.1", "_discovered": true }, { "name": "mlx-community/SongGeneration-v2-medium-4bit", "provider": "mlx-community", "parameter_count": "438.302M", "parameters_raw": 438301536, "min_ram_gb": 1.3, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-audio", "architecture": "", "hf_downloads": 24, "hf_likes": 0, "release_date": "2026-05-31", "format": "mlx", "mlx_only": true, "collection": "SongGeneration v2 MLX", "description": "Apple MLX checkpoints for Tencent SongGeneration v2 medium and large audiolm token generation.", "_discovered": true }, { "name": "mlx-community/gemma-3n-E4B-it-5bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.9, "recommended_ram_gb": 5.4, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 24, "hf_likes": 0, "release_date": "2025-07-12", "format": "mlx", "mlx_only": true, "collection": "Gemma 3n", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-TTS-12Hz-0.6B-Base-5bit", "provider": "mlx-community", "parameter_count": "600M", "parameters_raw": 600000000, "min_ram_gb": 1.4, "recommended_ram_gb": 2.5, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 23, "hf_likes": 1, "release_date": "2026-01-25", "format": "mlx", "mlx_only": true, "collection": "Qwen3-TTS", "description": "", "_discovered": true }, { "name": "mlx-community/Mellum-4b-sft-python", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 23, "hf_likes": 1, "release_date": "2025-06-28", "format": "mlx", "mlx_only": true, "collection": "JetBrains Mellum", "description": "Series of code models by JetBrains", "_discovered": true }, { "name": "mlx-community/olmOCR-7B-0225-preview-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 23, "hf_likes": 1, "release_date": "2025-03-03", "format": "mlx", "mlx_only": true, "collection": "olmOCR", "description": "", "_discovered": true }, { "name": "mlx-community/PaddleOCR-VL-6bit", "provider": "mlx-community", "parameter_count": "303.565M", "parameters_raw": 303564748, "min_ram_gb": 1.3, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 23, "hf_likes": 0, "release_date": "2026-01-19", "format": "mlx", "mlx_only": true, "collection": "PaddleOCR-VL", "description": "", "_discovered": true }, { "name": "mlx-community/K-EXAONE-236B-A23B-5bit", "provider": "mlx-community", "parameter_count": "236B", "parameters_raw": 236000000000, "min_ram_gb": 170.6, "recommended_ram_gb": 201.1, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 23, "hf_likes": 0, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "K-EXAONE", "description": " A large-scale multilingual language model by LG AI Research", "_discovered": true }, { "name": "mlx-community/Llama-OuteTTS-1.0-1B-fp16", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 22, "hf_likes": 3, "release_date": "2025-05-19", "format": "mlx", "mlx_only": true, "collection": "OuteTTS-1.0", "description": "", "_discovered": true }, { "name": "mlx-community/K-EXAONE-236B-A23B-6bit", "provider": "mlx-community", "parameter_count": "236B", "parameters_raw": 236000000000, "min_ram_gb": 204.5, "recommended_ram_gb": 241.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 22, "hf_likes": 0, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "K-EXAONE", "description": " A large-scale multilingual language model by LG AI Research", "_discovered": true }, { "name": "mlx-community/Nanonets-OCR-s-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 22, "hf_likes": 0, "release_date": "2025-07-25", "format": "mlx", "mlx_only": true, "collection": "Nanonets OCR", "description": "This collection houses Nanonets-OCR-s", "_discovered": true }, { "name": "mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-bf16", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 70.0, "recommended_ram_gb": 83.0, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 22, "hf_likes": 0, "release_date": "2025-06-19", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Qwen3", "description": "Abliterated, and further fine-tuned to be the most uncensored models available. Now in MLX", "_discovered": true }, { "name": "mlx-community/gemma-2-27b-4bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 16.5, "recommended_ram_gb": 20.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 22, "hf_likes": 0, "release_date": "2024-06-27", "format": "mlx", "mlx_only": true, "collection": "Google Gemma2", "description": "", "_discovered": true }, { "name": "mlx-community/QVQ-72B-Preview-4bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 42.4, "recommended_ram_gb": 50.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 21, "hf_likes": 7, "release_date": "2025-04-12", "format": "mlx", "mlx_only": true, "collection": "QVQ-72B-Preview", "description": "", "_discovered": true }, { "name": "mlx-community/Ling-2.6-flash-mlx-6bit", "provider": "mlx-community", "parameter_count": "104.187B", "parameters_raw": 104186907648, "min_ram_gb": 90.9, "recommended_ram_gb": 107.5, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 21, "hf_likes": 1, "release_date": "2026-04-29", "format": "mlx", "mlx_only": true, "collection": "inclusionAI LING 2.6", "description": "By inclusionAI", "_discovered": true }, { "name": "mlx-community/ERNIE-4.5-21B-A3B-PT-bf16", "provider": "mlx-community", "parameter_count": "21B", "parameters_raw": 21000000000, "min_ram_gb": 49.3, "recommended_ram_gb": 58.7, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 21, "hf_likes": 1, "release_date": "2025-07-04", "format": "mlx", "mlx_only": true, "collection": "ERNIE-4.5", "description": "", "_discovered": true }, { "name": "mlx-community/answerdotai-ModernBERT-base-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "fill-mask", "architecture": "", "hf_downloads": 21, "hf_likes": 1, "release_date": "2025-04-02", "format": "mlx", "mlx_only": true, "collection": "ModernBert", "description": "", "_discovered": true }, { "name": "mlx-community/paligemma2-3b-ft-docci-448-bf16", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 21, "hf_likes": 1, "release_date": "2024-12-05", "format": "mlx", "mlx_only": true, "collection": "Paligemma 2", "description": "", "_discovered": true }, { "name": "mlx-community/Molmo-7B-D-0924-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 21, "hf_likes": 1, "release_date": "2024-12-27", "format": "mlx", "mlx_only": true, "collection": "Molmo", "description": "", "_discovered": true }, { "name": "mlx-community/Josiefied-Olmo-3-7B-Instruct-abliterated-v1-bfloat16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 21, "hf_likes": 0, "release_date": "2025-11-25", "format": "mlx", "mlx_only": true, "collection": "Josiefied and Abliterated Models", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-VL-4B-Instruct-6bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 21, "hf_likes": 0, "release_date": "2025-10-14", "format": "mlx", "mlx_only": true, "collection": "Qwen3-VL", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen2.5-32B-Instruct-bf16", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 74.6, "recommended_ram_gb": 88.4, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 21, "hf_likes": 0, "release_date": "2024-09-18", "format": "mlx", "mlx_only": true, "collection": "Qwen2.5", "description": "The Qwen 2.5 models are a series of AI models trained on 18 trillion tokens, supporting 29 languages and offering advanced features such as instructio", "_discovered": true }, { "name": "mlx-community/INTELLECT-3-4bit", "provider": "mlx-community", "parameter_count": "106.852B", "parameters_raw": 106852251264, "min_ram_gb": 62.4, "recommended_ram_gb": 74.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 20, "hf_likes": 3, "release_date": "2025-11-27", "format": "mlx", "mlx_only": true, "collection": "INTELLECT 3", "description": "", "_discovered": true }, { "name": "mlx-community/LFM2-8B-A1B-6bit-MLX", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 20, "hf_likes": 3, "release_date": "2025-10-08", "format": "mlx", "mlx_only": true, "collection": "💧LFM2-8B-A1B-MoE", "description": "Best in Class MoE, better than Qwen3. Optimised for Smaller devices sub 16 GB (M1/2/3/4) Apple Silicon.", "_discovered": true }, { "name": "mlx-community/Qwen1.5-7B-Chat-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 20, "hf_likes": 2, "release_date": "2024-03-07", "format": "mlx", "mlx_only": true, "collection": "Qwen1.5", "description": "Qwen1.5 is the improved version of Qwen, the large language model series developed by Alibaba Cloud.", "_discovered": true }, { "name": "mlx-community/NAFNet-REDS-width64", "provider": "mlx-community", "parameter_count": "67.8888M", "parameters_raw": 67888835, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 20, "hf_likes": 1, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "NAFNet MLX", "description": "MLX port of NAFNet (Simple Baselines for Image Restoration): on-device deblur/denoise on Apple Silicon.", "_discovered": true }, { "name": "mlx-community/Ling-2.6-flash-mlx-8bit", "provider": "mlx-community", "parameter_count": "104.187B", "parameters_raw": 104186907648, "min_ram_gb": 120.8, "recommended_ram_gb": 142.7, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 20, "hf_likes": 1, "release_date": "2026-04-29", "format": "mlx", "mlx_only": true, "collection": "inclusionAI LING 2.6", "description": "By inclusionAI", "_discovered": true }, { "name": "mlx-community/Olmo-3-7B-Instruct-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 20, "hf_likes": 1, "release_date": "2025-11-20", "format": "mlx", "mlx_only": true, "collection": "Olmo-3", "description": "Ai2's Olmo 3 model family of instruction and reasoning models.", "_discovered": true }, { "name": "mlx-community/VisualQuality-R1-7B-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "reinforcement-learning", "architecture": "", "hf_downloads": 20, "hf_likes": 1, "release_date": "2025-08-06", "format": "mlx", "mlx_only": true, "collection": "VisualQuality-R1", "description": "Image Quality Assessment", "_discovered": true }, { "name": "mlx-community/P1-VL-30B-A3B-bf16", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 70.0, "recommended_ram_gb": 83.0, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 20, "hf_likes": 0, "release_date": "2026-05-18", "format": "mlx", "mlx_only": true, "collection": "PRIME-RL P1-VL-30B-A3B", "description": "Bridging visual perception and scientific reasoning in physics olympiads", "_discovered": true }, { "name": "mlx-community/PaddleOCR-VL-5bit", "provider": "mlx-community", "parameter_count": "279.483M", "parameters_raw": 279483272, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 20, "hf_likes": 0, "release_date": "2026-01-19", "format": "mlx", "mlx_only": true, "collection": "PaddleOCR-VL", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-4B-Instruct-2507-gabliterated-4bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 20, "hf_likes": 0, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "Gabliterated v1", "description": "The next version of Abliteration", "_discovered": true }, { "name": "mlx-community/Qwen3-VL-4B-Instruct-5bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.9, "recommended_ram_gb": 5.4, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 20, "hf_likes": 0, "release_date": "2025-10-14", "format": "mlx", "mlx_only": true, "collection": "Qwen3-VL", "description": "", "_discovered": true }, { "name": "mlx-community/LFM2-1.2B-5bit", "provider": "mlx-community", "parameter_count": "1.2B", "parameters_raw": 1200000000, "min_ram_gb": 1.9, "recommended_ram_gb": 3.0, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 19, "hf_likes": 1, "release_date": "2025-07-12", "format": "mlx", "mlx_only": true, "collection": "LFM2.x", "description": "", "_discovered": true }, { "name": "mlx-community/Mistral-Nemo-Base-2407-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 19, "hf_likes": 1, "release_date": "2024-07-18", "format": "mlx", "mlx_only": true, "collection": "Mistral NeMo", "description": "", "_discovered": true }, { "name": "mlx-community/SmolLM-135M-fp16", "provider": "mlx-community", "parameter_count": "135M", "parameters_raw": 135000000, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 19, "hf_likes": 1, "release_date": "2024-07-16", "format": "mlx", "mlx_only": true, "collection": "HF SmolLM", "description": "A series of smol LLMs: 135M, 360M and 1.7B.", "_discovered": true }, { "name": "mlx-community/SU-01-bf16", "provider": "mlx-community", "parameter_count": "30.5321B", "parameters_raw": 30532122624, "min_ram_gb": 71.2, "recommended_ram_gb": 84.4, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 19, "hf_likes": 0, "release_date": "2026-05-16", "format": "mlx", "mlx_only": true, "collection": "Simplified Reasoning SU-01", "description": "Rigorous mathematical and scientific olympiad problem solving", "_discovered": true }, { "name": "mlx-community/Gemma-SEA-LION-v3-9B-IT-mlx-4bit", "provider": "mlx-community", "parameter_count": "9B", "parameters_raw": 9000000000, "min_ram_gb": 6.2, "recommended_ram_gb": 8.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 19, "hf_likes": 0, "release_date": "2025-09-10", "format": "mlx", "mlx_only": true, "collection": "SEA-LION", "description": "SEA-LION mlx models by AI Singapore.", "_discovered": true }, { "name": "mlx-community/SmolLM-360M-8bit", "provider": "mlx-community", "parameter_count": "360M", "parameters_raw": 360000000, "min_ram_gb": 1.4, "recommended_ram_gb": 2.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 19, "hf_likes": 0, "release_date": "2024-07-16", "format": "mlx", "mlx_only": true, "collection": "HF SmolLM", "description": "A series of smol LLMs: 135M, 360M and 1.7B.", "_discovered": true }, { "name": "mlx-community/Dia-1.6B-3bit", "provider": "mlx-community", "parameter_count": "1.6B", "parameters_raw": 1600000000, "min_ram_gb": 1.7, "recommended_ram_gb": 2.8, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 18, "hf_likes": 4, "release_date": "2025-04-24", "format": "mlx", "mlx_only": true, "collection": "NariLabs Dia-1.5B", "description": "", "_discovered": true }, { "name": "mlx-community/DR-Venus-4B-SFT-mlx-8Bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 18, "hf_likes": 2, "release_date": "2026-04-29", "format": "mlx", "mlx_only": true, "collection": "inclusionAI DR-Venus", "description": "By inclusionAI", "_discovered": true }, { "name": "mlx-community/functiongemma-270m-it-8bit", "provider": "mlx-community", "parameter_count": "270M", "parameters_raw": 270000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 18, "hf_likes": 1, "release_date": "2025-12-18", "format": "mlx", "mlx_only": true, "collection": "FunctionGemma", "description": "by Google Deepmind", "_discovered": true }, { "name": "mlx-community/mamba-1.4b-hf-f16", "provider": "mlx-community", "parameter_count": "1.4B", "parameters_raw": 1400000000, "min_ram_gb": 1.8, "recommended_ram_gb": 2.9, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 18, "hf_likes": 1, "release_date": "2024-09-21", "format": "mlx", "mlx_only": true, "collection": "Mamba", "description": "Mamba is a new LLM architecture that integrates the Structured State Space sequence model to manage lengthy data sequences.", "_discovered": true }, { "name": "mlx-community/SongGeneration-v2-medium-fp32", "provider": "mlx-community", "parameter_count": "2.80442B", "parameters_raw": 2804416512, "min_ram_gb": 2.6, "recommended_ram_gb": 3.9, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-audio", "architecture": "", "hf_downloads": 18, "hf_likes": 0, "release_date": "2026-05-31", "format": "mlx", "mlx_only": true, "collection": "SongGeneration v2 MLX", "description": "Apple MLX checkpoints for Tencent SongGeneration v2 medium and large audiolm token generation.", "_discovered": true }, { "name": "mlx-community/SU-01-8bit", "provider": "mlx-community", "parameter_count": "30.5321B", "parameters_raw": 30532122624, "min_ram_gb": 36.1, "recommended_ram_gb": 43.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 18, "hf_likes": 0, "release_date": "2026-05-17", "format": "mlx", "mlx_only": true, "collection": "Simplified Reasoning SU-01", "description": "Rigorous mathematical and scientific olympiad problem solving", "_discovered": true }, { "name": "mlx-community/AceReason-Nemotron-14B-4bit", "provider": "mlx-community", "parameter_count": "14B", "parameters_raw": 14000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 18, "hf_likes": 0, "release_date": "2025-05-24", "format": "mlx", "mlx_only": true, "collection": "AceReason Nemotron", "description": "", "_discovered": true }, { "name": "mlx-community/GLM-4-32B-Base-0414-8bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 37.8, "recommended_ram_gb": 45.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 18, "hf_likes": 0, "release_date": "2025-04-21", "format": "mlx", "mlx_only": true, "collection": "GLM4", "description": "The GLM-4 and Z1 series are powerful open-source language models excelling in reasoning, code, and complex tasks.", "_discovered": true }, { "name": "mlx-community/QwQ-32B-Preview-3bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 14.8, "recommended_ram_gb": 18.2, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 17, "hf_likes": 5, "release_date": "2024-11-28", "format": "mlx", "mlx_only": true, "collection": "QwQ-32B-Preview", "description": "", "_discovered": true }, { "name": "mlx-community/QwQ-32B-Preview-4bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 19.4, "recommended_ram_gb": 23.6, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 17, "hf_likes": 3, "release_date": "2024-11-27", "format": "mlx", "mlx_only": true, "collection": "QwQ-32B-Preview", "description": "", "_discovered": true }, { "name": "mlx-community/IQuest-Coder-V1-40B-Instruct-6bit", "provider": "mlx-community", "parameter_count": "40B", "parameters_raw": 40000000000, "min_ram_gb": 35.5, "recommended_ram_gb": 42.5, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 17, "hf_likes": 1, "release_date": "2026-01-01", "format": "mlx", "mlx_only": true, "collection": "IQuest-Coder", "description": "By IQuestLab", "_discovered": true }, { "name": "mlx-community/VibeVoice-Realtime-0.5B-6bit", "provider": "mlx-community", "parameter_count": "500M", "parameters_raw": 500000000, "min_ram_gb": 1.4, "recommended_ram_gb": 2.5, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 17, "hf_likes": 1, "release_date": "2025-12-15", "format": "mlx", "mlx_only": true, "collection": "VibeVoice", "description": "", "_discovered": true }, { "name": "mlx-community/Holo1-3B-4bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 17, "hf_likes": 1, "release_date": "2025-06-03", "format": "mlx", "mlx_only": true, "collection": "Holo1", "description": "", "_discovered": true }, { "name": "mlx-community/Molmo-7B-D-0924-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 17, "hf_likes": 1, "release_date": "2025-01-01", "format": "mlx", "mlx_only": true, "collection": "Molmo", "description": "", "_discovered": true }, { "name": "mlx-community/LLaDA2.0-mini-8bit", "provider": "mlx-community", "parameter_count": "16.2556B", "parameters_raw": 16255643392, "min_ram_gb": 19.7, "recommended_ram_gb": 23.9, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 17, "hf_likes": 0, "release_date": "2025-11-26", "format": "mlx", "mlx_only": true, "collection": "LLaDA 2.0", "description": "", "_discovered": true }, { "name": "mlx-community/Nanonets-OCR-s-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 17, "hf_likes": 0, "release_date": "2025-07-25", "format": "mlx", "mlx_only": true, "collection": "Nanonets OCR", "description": "This collection houses Nanonets-OCR-s", "_discovered": true }, { "name": "mlx-community/gemma-3-27b-it-qat-6bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 24.3, "recommended_ram_gb": 29.3, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 17, "hf_likes": 0, "release_date": "2025-04-19", "format": "mlx", "mlx_only": true, "collection": "Gemma 3 QAT", "description": "Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory.", "_discovered": true }, { "name": "mlx-community/SmolLM-360M-4bit", "provider": "mlx-community", "parameter_count": "360M", "parameters_raw": 360000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 17, "hf_likes": 0, "release_date": "2024-07-16", "format": "mlx", "mlx_only": true, "collection": "HF SmolLM", "description": "A series of smol LLMs: 135M, 360M and 1.7B.", "_discovered": true }, { "name": "mlx-community/DR-Venus-4B-RL-mlx-8Bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 16, "hf_likes": 2, "release_date": "2026-04-29", "format": "mlx", "mlx_only": true, "collection": "inclusionAI DR-Venus", "description": "By inclusionAI", "_discovered": true }, { "name": "mlx-community/ERNIE-4.5-21B-A3B-PT-8bit", "provider": "mlx-community", "parameter_count": "21B", "parameters_raw": 21000000000, "min_ram_gb": 25.1, "recommended_ram_gb": 30.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 16, "hf_likes": 2, "release_date": "2025-07-04", "format": "mlx", "mlx_only": true, "collection": "ERNIE-4.5", "description": "", "_discovered": true }, { "name": "mlx-community/YOLO26l-OptiQ-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "object-detection", "architecture": "", "hf_downloads": 16, "hf_likes": 0, "release_date": "2026-04-26", "format": "mlx", "mlx_only": true, "collection": "YOLO 26", "description": "", "_discovered": true }, { "name": "mlx-community/kitten-tts-nano-0.8-4bit", "provider": "mlx-community", "parameter_count": "7.55015M", "parameters_raw": 7550146, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 16, "hf_likes": 0, "release_date": "2026-02-24", "format": "mlx", "mlx_only": true, "collection": "KittenTTS", "description": "All MLX conversions of KittenTTS (nano/micro/mini) across fp32, fp16, bf16, and 4/5/6/8-bit quantizations.", "_discovered": true }, { "name": "mlx-community/olmOCR-2-7B-1025-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 16, "hf_likes": 0, "release_date": "2025-10-25", "format": "mlx", "mlx_only": true, "collection": "olmOCR 2", "description": "", "_discovered": true }, { "name": "mlx-community/whisper-tiny.en-mlx-q4", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 16, "hf_likes": 0, "release_date": "2024-03-09", "format": "mlx", "mlx_only": true, "collection": "Whisper", "description": "OpenAI Whisper speech recognition models in MLX format", "_discovered": true }, { "name": "mlx-community/Apriel-1.5-15b-Thinker-6bit-MLX", "provider": "mlx-community", "parameter_count": "15B", "parameters_raw": 15000000000, "min_ram_gb": 13.9, "recommended_ram_gb": 17.2, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 15, "hf_likes": 1, "release_date": "2025-10-03", "format": "mlx", "mlx_only": true, "collection": "ServiceNow-Apriel", "description": "Apriel-1.5-15b-Thinker is a multimodal reasoning model in ServiceNow’s Apriel SLM series which achieves competitive performance against models 10 time", "_discovered": true }, { "name": "mlx-community/chatterbox-turbo-6bit", "provider": "mlx-community", "parameter_count": "170.912M", "parameters_raw": 170912322, "min_ram_gb": 1.1, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 15, "hf_likes": 0, "release_date": "2025-12-17", "format": "mlx", "mlx_only": true, "collection": "Chatterbox TTS", "description": "Chatterbox and Chatterbox Turbo By ResembleAI", "_discovered": true }, { "name": "mlx-community/Apriel-1.5-15b-Thinker-3bit-MLX", "provider": "mlx-community", "parameter_count": "15B", "parameters_raw": 15000000000, "min_ram_gb": 7.5, "recommended_ram_gb": 9.6, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 15, "hf_likes": 0, "release_date": "2025-10-03", "format": "mlx", "mlx_only": true, "collection": "ServiceNow-Apriel", "description": "Apriel-1.5-15b-Thinker is a multimodal reasoning model in ServiceNow’s Apriel SLM series which achieves competitive performance against models 10 time", "_discovered": true }, { "name": "mlx-community/LFM2-1.2B-6bit", "provider": "mlx-community", "parameter_count": "1.2B", "parameters_raw": 1200000000, "min_ram_gb": 2.0, "recommended_ram_gb": 3.2, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 14, "hf_likes": 3, "release_date": "2025-07-12", "format": "mlx", "mlx_only": true, "collection": "LFM2.x", "description": "", "_discovered": true }, { "name": "mlx-community/swahili-gemma-1b-mlx-fp16", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 14, "hf_likes": 1, "release_date": "2025-08-26", "format": "mlx", "mlx_only": true, "collection": "Swahili Gemma 1B", "description": "A fine-tuned Gemma 3 1B instruction model specialized for English-to-Swahili translation and Swahili conversational AI. The model accepts input in bot", "_discovered": true }, { "name": "mlx-community/SoulX-Singer-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-audio", "architecture": "", "hf_downloads": 14, "hf_likes": 0, "release_date": "2026-05-31", "format": "mlx", "mlx_only": true, "collection": "SoulX-Singer MLX", "description": "Apple MLX safetensors checkpoints for Soul-AILab SoulX-Singer and SoulX-Singer-SVC.", "_discovered": true }, { "name": "mlx-community/SongGeneration-v2-medium-bf16", "provider": "mlx-community", "parameter_count": "2.80442B", "parameters_raw": 2804416512, "min_ram_gb": 7.5, "recommended_ram_gb": 9.6, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-audio", "architecture": "", "hf_downloads": 14, "hf_likes": 0, "release_date": "2026-05-31", "format": "mlx", "mlx_only": true, "collection": "SongGeneration v2 MLX", "description": "Apple MLX checkpoints for Tencent SongGeneration v2 medium and large audiolm token generation.", "_discovered": true }, { "name": "mlx-community/VibeVoice-Realtime-0.5B-5bit", "provider": "mlx-community", "parameter_count": "500M", "parameters_raw": 500000000, "min_ram_gb": 1.4, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 14, "hf_likes": 0, "release_date": "2025-12-15", "format": "mlx", "mlx_only": true, "collection": "VibeVoice", "description": "", "_discovered": true }, { "name": "mlx-community/Nanonets-OCR2-3B-6bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 3.6, "recommended_ram_gb": 5.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 14, "hf_likes": 0, "release_date": "2025-10-14", "format": "mlx", "mlx_only": true, "collection": "Nanonets OCR2", "description": "This collection houses Nanonets-OCR2 models", "_discovered": true }, { "name": "mlx-community/UI-TARS-7B-SFT-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 14, "hf_likes": 0, "release_date": "2025-03-03", "format": "mlx", "mlx_only": true, "collection": "UI-TARS", "description": "", "_discovered": true }, { "name": "mlx-community/sam-audio-small-fp16", "provider": "mlx-community", "parameter_count": "602.312M", "parameters_raw": 602312324, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 13, "hf_likes": 1, "release_date": "2025-12-23", "format": "mlx", "mlx_only": true, "collection": "Sam Audio", "description": "By Facebook ", "_discovered": true }, { "name": "mlx-community/Yi-1.5-9B-8bit", "provider": "mlx-community", "parameter_count": "9B", "parameters_raw": 9000000000, "min_ram_gb": 11.3, "recommended_ram_gb": 14.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 13, "hf_likes": 1, "release_date": "2024-05-13", "format": "mlx", "mlx_only": true, "collection": "Yi-1.5", "description": "", "_discovered": true }, { "name": "mlx-community/IQuest-Coder-V1-40B-Instruct-5bit", "provider": "mlx-community", "parameter_count": "40B", "parameters_raw": 40000000000, "min_ram_gb": 29.7, "recommended_ram_gb": 35.8, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 13, "hf_likes": 0, "release_date": "2026-01-02", "format": "mlx", "mlx_only": true, "collection": "IQuest-Coder", "description": "By IQuestLab", "_discovered": true }, { "name": "mlx-community/granite-4.0-h-1b-3bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.4, "recommended_ram_gb": 2.5, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 13, "hf_likes": 0, "release_date": "2025-10-28", "format": "mlx", "mlx_only": true, "collection": "Granite 4.0 Nano Language Models", "description": "", "_discovered": true }, { "name": "mlx-community/OLMoE-1B-7B-0125-Instruct-8bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 13, "hf_likes": 0, "release_date": "2025-03-04", "format": "mlx", "mlx_only": true, "collection": "OLMoE", "description": "", "_discovered": true }, { "name": "mlx-community/olmOCR-7B-0225-preview-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 13, "hf_likes": 0, "release_date": "2025-03-03", "format": "mlx", "mlx_only": true, "collection": "olmOCR", "description": "", "_discovered": true }, { "name": "mlx-community/paligemma2-3b-ft-docci-448-8bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 13, "hf_likes": 0, "release_date": "2024-12-05", "format": "mlx", "mlx_only": true, "collection": "Paligemma 2", "description": "", "_discovered": true }, { "name": "mlx-community/SmolLM-135M-8bit", "provider": "mlx-community", "parameter_count": "135M", "parameters_raw": 135000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 13, "hf_likes": 0, "release_date": "2024-07-16", "format": "mlx", "mlx_only": true, "collection": "HF SmolLM", "description": "A series of smol LLMs: 135M, 360M and 1.7B.", "_discovered": true }, { "name": "mlx-community/OpenELM-3B", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 12, "hf_likes": 8, "release_date": "2024-04-26", "format": "mlx", "mlx_only": true, "collection": "OpenELM", "description": "A family of Open-source Efficient Language Models from Apple.", "_discovered": true }, { "name": "mlx-community/QwQ-32B-Preview-8bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 37.8, "recommended_ram_gb": 45.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 12, "hf_likes": 6, "release_date": "2024-11-27", "format": "mlx", "mlx_only": true, "collection": "QwQ-32B-Preview", "description": "", "_discovered": true }, { "name": "mlx-community/QwQ-32B-Preview-6bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 28.6, "recommended_ram_gb": 34.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 12, "hf_likes": 4, "release_date": "2024-11-27", "format": "mlx", "mlx_only": true, "collection": "QwQ-32B-Preview", "description": "", "_discovered": true }, { "name": "mlx-community/Orchestrator-8B-6bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 7.9, "recommended_ram_gb": 10.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 12, "hf_likes": 1, "release_date": "2025-11-29", "format": "mlx", "mlx_only": true, "collection": "Orchestrator 8B", "description": "", "_discovered": true }, { "name": "mlx-community/Holo1-3B-3bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.3, "recommended_ram_gb": 3.5, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 12, "hf_likes": 1, "release_date": "2025-06-03", "format": "mlx", "mlx_only": true, "collection": "Holo1", "description": "", "_discovered": true }, { "name": "mlx-community/answerdotai-ModernBERT-base-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "fill-mask", "architecture": "", "hf_downloads": 12, "hf_likes": 1, "release_date": "2025-04-02", "format": "mlx", "mlx_only": true, "collection": "ModernBert", "description": "", "_discovered": true }, { "name": "mlx-community/SU-01-6bit", "provider": "mlx-community", "parameter_count": "30.5321B", "parameters_raw": 30532122624, "min_ram_gb": 27.3, "recommended_ram_gb": 32.9, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 12, "hf_likes": 0, "release_date": "2026-05-16", "format": "mlx", "mlx_only": true, "collection": "Simplified Reasoning SU-01", "description": "Rigorous mathematical and scientific olympiad problem solving", "_discovered": true }, { "name": "mlx-community/INTELLECT-3-6bit", "provider": "mlx-community", "parameter_count": "106.852B", "parameters_raw": 106852251264, "min_ram_gb": 93.2, "recommended_ram_gb": 110.2, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 12, "hf_likes": 0, "release_date": "2025-11-27", "format": "mlx", "mlx_only": true, "collection": "INTELLECT 3", "description": "", "_discovered": true }, { "name": "mlx-community/olmOCR-7B-0725-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 12, "hf_likes": 0, "release_date": "2025-07-24", "format": "mlx", "mlx_only": true, "collection": "olmOCR-0725", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3n-E2B-it-5bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.4, "recommended_ram_gb": 3.7, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 12, "hf_likes": 0, "release_date": "2025-07-12", "format": "mlx", "mlx_only": true, "collection": "Gemma 3n", "description": "", "_discovered": true }, { "name": "mlx-community/Mistral-Small-24B-Instruct-2501-6bit", "provider": "mlx-community", "parameter_count": "24B", "parameters_raw": 24000000000, "min_ram_gb": 21.7, "recommended_ram_gb": 26.3, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 12, "hf_likes": 0, "release_date": "2025-01-30", "format": "mlx", "mlx_only": true, "collection": "Mistral Small", "description": "", "_discovered": true }, { "name": "mlx-community/QVQ-72B-Preview-8bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 83.8, "recommended_ram_gb": 99.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 11, "hf_likes": 3, "release_date": "2024-12-24", "format": "mlx", "mlx_only": true, "collection": "QVQ-72B-Preview", "description": "", "_discovered": true }, { "name": "mlx-community/NAFNet-SIDD-width64", "provider": "mlx-community", "parameter_count": "115.983M", "parameters_raw": 115982915, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 11, "hf_likes": 1, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "NAFNet MLX", "description": "MLX port of NAFNet (Simple Baselines for Image Restoration): on-device deblur/denoise on Apple Silicon.", "_discovered": true }, { "name": "mlx-community/Gemma-SEA-LION-v4-27B-IT-mlx-4bit", "provider": "mlx-community", "parameter_count": "27B", "parameters_raw": 27000000000, "min_ram_gb": 16.5, "recommended_ram_gb": 20.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 11, "hf_likes": 1, "release_date": "2025-09-10", "format": "mlx", "mlx_only": true, "collection": "SEA-LION", "description": "SEA-LION mlx models by AI Singapore.", "_discovered": true }, { "name": "mlx-community/VoxCPM1.5-6bit", "provider": "mlx-community", "parameter_count": "261.525M", "parameters_raw": 261525441, "min_ram_gb": 1.2, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 11, "hf_likes": 0, "release_date": "2025-12-16", "format": "mlx", "mlx_only": true, "collection": "VoxCPM", "description": "", "_discovered": true }, { "name": "mlx-community/GLM-4-32B-Base-0414-6bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 28.6, "recommended_ram_gb": 34.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 11, "hf_likes": 0, "release_date": "2025-04-21", "format": "mlx", "mlx_only": true, "collection": "GLM4", "description": "The GLM-4 and Z1 series are powerful open-source language models excelling in reasoning, code, and complex tasks.", "_discovered": true }, { "name": "mlx-community/Virtuoso-Medium-v2-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 11, "hf_likes": 0, "release_date": "2025-01-30", "format": "mlx", "mlx_only": true, "collection": "Arcee Virtuoso", "description": "", "_discovered": true }, { "name": "mlx-community/falcon-mamba-7b-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 11, "hf_likes": 0, "release_date": "2024-11-15", "format": "mlx", "mlx_only": true, "collection": "Falcon-Mamba", "description": "Falcon Mamba models compatible with MLX", "_discovered": true }, { "name": "mlx-community/SmolLM3-3B-3bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.3, "recommended_ram_gb": 3.5, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 10, "hf_likes": 2, "release_date": "2025-07-08", "format": "mlx", "mlx_only": true, "collection": "SmolLM3", "description": "", "_discovered": true }, { "name": "mlx-community/LFM2-VL-450M-5bit", "provider": "mlx-community", "parameter_count": "450M", "parameters_raw": 450000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.4, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 10, "hf_likes": 1, "release_date": "2025-08-16", "format": "mlx", "mlx_only": true, "collection": "LFM2-VL", "description": "", "_discovered": true }, { "name": "mlx-community/helium-1-preview-2b-4bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 10, "hf_likes": 1, "release_date": "2025-01-18", "format": "mlx", "mlx_only": true, "collection": "Helium-1", "description": "Kyutai's Helium-1 2B Model, outperforming other state of the art small models.", "_discovered": true }, { "name": "mlx-community/Qwen3-4B-Instruct-2507-gabliterated-mxfp4", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 10, "hf_likes": 0, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "Gabliterated v1", "description": "The next version of Abliteration", "_discovered": true }, { "name": "mlx-community/chatterbox-turbo-5bit", "provider": "mlx-community", "parameter_count": "152.71M", "parameters_raw": 152709762, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 10, "hf_likes": 0, "release_date": "2025-12-17", "format": "mlx", "mlx_only": true, "collection": "Chatterbox TTS", "description": "Chatterbox and Chatterbox Turbo By ResembleAI", "_discovered": true }, { "name": "mlx-community/gemma-3-270m-it-5bit", "provider": "mlx-community", "parameter_count": "270M", "parameters_raw": 270000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 10, "hf_likes": 0, "release_date": "2025-08-09", "format": "mlx", "mlx_only": true, "collection": "Gemma 3-270m", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-270m-it-6bit", "provider": "mlx-community", "parameter_count": "270M", "parameters_raw": 270000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 10, "hf_likes": 0, "release_date": "2025-08-09", "format": "mlx", "mlx_only": true, "collection": "Gemma 3-270m", "description": "", "_discovered": true }, { "name": "mlx-community/olmOCR-7B-0725-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 10, "hf_likes": 0, "release_date": "2025-07-25", "format": "mlx", "mlx_only": true, "collection": "olmOCR-0725", "description": "", "_discovered": true }, { "name": "mlx-community/OLMoE-1B-7B-0125-Instruct", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 10, "hf_likes": 0, "release_date": "2025-03-04", "format": "mlx", "mlx_only": true, "collection": "OLMoE", "description": "", "_discovered": true }, { "name": "mlx-community/EXAONE-3.5-2.4B-Instruct-bf16", "provider": "mlx-community", "parameter_count": "2.4B", "parameters_raw": 2400000000, "min_ram_gb": 6.5, "recommended_ram_gb": 8.5, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 10, "hf_likes": 0, "release_date": "2024-12-09", "format": "mlx", "mlx_only": true, "collection": "EXAONE-3.5", "description": "EXAONE 3.5, a collection of instruction-tuned bilingual generative models ranging from 2.4B to 32B parameters, developed by LG AI.", "_discovered": true }, { "name": "mlx-community/falcon-mamba-7b-4bit-instruct", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 10, "hf_likes": 0, "release_date": "2024-11-15", "format": "mlx", "mlx_only": true, "collection": "Falcon-Mamba", "description": "Falcon Mamba models compatible with MLX", "_discovered": true }, { "name": "mlx-community/Olmo-3-7B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 9, "hf_likes": 2, "release_date": "2025-11-20", "format": "mlx", "mlx_only": true, "collection": "Olmo-3", "description": "Ai2's Olmo 3 model family of instruction and reasoning models.", "_discovered": true }, { "name": "mlx-community/kitten-tts-nano-0.8-5bit", "provider": "mlx-community", "parameter_count": "7.81093M", "parameters_raw": 7810930, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 9, "hf_likes": 1, "release_date": "2026-02-24", "format": "mlx", "mlx_only": true, "collection": "KittenTTS", "description": "All MLX conversions of KittenTTS (nano/micro/mini) across fp32, fp16, bf16, and 4/5/6/8-bit quantizations.", "_discovered": true }, { "name": "mlx-community/Virtuoso-Medium-v2-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 9, "hf_likes": 1, "release_date": "2025-01-31", "format": "mlx", "mlx_only": true, "collection": "Arcee Virtuoso", "description": "", "_discovered": true }, { "name": "mlx-community/Qwen3-4B-Instruct-2507-gabliterated-6bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 9, "hf_likes": 0, "release_date": "2026-01-06", "format": "mlx", "mlx_only": true, "collection": "Gabliterated v1", "description": "The next version of Abliteration", "_discovered": true }, { "name": "mlx-community/VoxCPM1.5-5bit", "provider": "mlx-community", "parameter_count": "236.475M", "parameters_raw": 236475009, "min_ram_gb": 1.2, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 9, "hf_likes": 0, "release_date": "2025-12-16", "format": "mlx", "mlx_only": true, "collection": "VoxCPM", "description": "", "_discovered": true }, { "name": "mlx-community/LLaDA2.0-mini-6bit", "provider": "mlx-community", "parameter_count": "16.2556B", "parameters_raw": 16255643392, "min_ram_gb": 15.0, "recommended_ram_gb": 18.5, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 9, "hf_likes": 0, "release_date": "2025-11-26", "format": "mlx", "mlx_only": true, "collection": "LLaDA 2.0", "description": "", "_discovered": true }, { "name": "mlx-community/SmolLM3-3B-5bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 3.2, "recommended_ram_gb": 4.5, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 9, "hf_likes": 0, "release_date": "2025-07-08", "format": "mlx", "mlx_only": true, "collection": "SmolLM3", "description": "", "_discovered": true }, { "name": "mlx-community/Holo1-3B-6bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 3.6, "recommended_ram_gb": 5.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 9, "hf_likes": 0, "release_date": "2025-06-03", "format": "mlx", "mlx_only": true, "collection": "Holo1", "description": "", "_discovered": true }, { "name": "mlx-community/OLMoE-1B-7B-0125-Instruct-6bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.9, "recommended_ram_gb": 3.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 9, "hf_likes": 0, "release_date": "2025-03-04", "format": "mlx", "mlx_only": true, "collection": "OLMoE", "description": "", "_discovered": true }, { "name": "mlx-community/UI-TARS-7B-SFT-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 9, "hf_likes": 0, "release_date": "2025-03-03", "format": "mlx", "mlx_only": true, "collection": "UI-TARS", "description": "", "_discovered": true }, { "name": "mlx-community/QVQ-72B-Preview-3bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 32.0, "recommended_ram_gb": 38.5, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 8, "hf_likes": 5, "release_date": "2024-12-24", "format": "mlx", "mlx_only": true, "collection": "QVQ-72B-Preview", "description": "", "_discovered": true }, { "name": "mlx-community/QwQ-32B-6bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 28.6, "recommended_ram_gb": 34.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 8, "hf_likes": 3, "release_date": "2025-03-05", "format": "mlx", "mlx_only": true, "collection": "Qwen QwQ", "description": "", "_discovered": true }, { "name": "mlx-community/Llama-SEA-LION-v3.5-8B-R-mlx-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 8, "hf_likes": 1, "release_date": "2025-09-10", "format": "mlx", "mlx_only": true, "collection": "SEA-LION", "description": "SEA-LION mlx models by AI Singapore.", "_discovered": true }, { "name": "mlx-community/Orchestrator-8B-5bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 6.8, "recommended_ram_gb": 8.8, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 8, "hf_likes": 0, "release_date": "2025-11-29", "format": "mlx", "mlx_only": true, "collection": "Orchestrator 8B", "description": "", "_discovered": true }, { "name": "mlx-community/Llama-SEA-LION-v3-8B-IT-mlx-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 8, "hf_likes": 0, "release_date": "2025-09-10", "format": "mlx", "mlx_only": true, "collection": "SEA-LION", "description": "SEA-LION mlx models by AI Singapore.", "_discovered": true }, { "name": "mlx-community/olmOCR-7B-0725-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 8, "hf_likes": 0, "release_date": "2025-07-25", "format": "mlx", "mlx_only": true, "collection": "olmOCR-0725", "description": "", "_discovered": true }, { "name": "mlx-community/QwQ-32B-3bit", "provider": "mlx-community", "parameter_count": "32B", "parameters_raw": 32000000000, "min_ram_gb": 14.8, "recommended_ram_gb": 18.2, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 8, "hf_likes": 0, "release_date": "2025-03-05", "format": "mlx", "mlx_only": true, "collection": "Qwen QwQ", "description": "", "_discovered": true }, { "name": "mlx-community/encodec-48khz-float32", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7, "hf_likes": 3, "release_date": "2024-09-16", "format": "mlx", "mlx_only": true, "collection": "EnCodec", "description": "EnCodec models in MLX", "_discovered": true }, { "name": "mlx-community/INTELLECT-3-5bit", "provider": "mlx-community", "parameter_count": "106.852B", "parameters_raw": 106852251264, "min_ram_gb": 77.8, "recommended_ram_gb": 92.2, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7, "hf_likes": 1, "release_date": "2025-11-27", "format": "mlx", "mlx_only": true, "collection": "INTELLECT 3", "description": "", "_discovered": true }, { "name": "mlx-community/granite-4.0-h-1b-5bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.7, "recommended_ram_gb": 2.8, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7, "hf_likes": 1, "release_date": "2025-10-28", "format": "mlx", "mlx_only": true, "collection": "Granite 4.0 Nano Language Models", "description": "", "_discovered": true }, { "name": "mlx-community/helium-1-preview-2b-8bit", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 3.3, "recommended_ram_gb": 4.7, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7, "hf_likes": 1, "release_date": "2025-01-18", "format": "mlx", "mlx_only": true, "collection": "Helium-1", "description": "Kyutai's Helium-1 2B Model, outperforming other state of the art small models.", "_discovered": true }, { "name": "mlx-community/P1-VL-30B-A3B-5bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 22.6, "recommended_ram_gb": 27.3, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2026-05-18", "format": "mlx", "mlx_only": true, "collection": "PRIME-RL P1-VL-30B-A3B", "description": "Bridging visual perception and scientific reasoning in physics olympiads", "_discovered": true }, { "name": "mlx-community/kitten-tts-nano-0.8-6bit", "provider": "mlx-community", "parameter_count": "8.07171M", "parameters_raw": 8071714, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2026-02-24", "format": "mlx", "mlx_only": true, "collection": "KittenTTS", "description": "All MLX conversions of KittenTTS (nano/micro/mini) across fp32, fp16, bf16, and 4/5/6/8-bit quantizations.", "_discovered": true }, { "name": "mlx-community/PE-Core-G14-448", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2025-12-26", "format": "mlx", "mlx_only": true, "collection": "Perception Encoder", "description": "Perception Encoder Models from Facebook", "_discovered": true }, { "name": "mlx-community/VisualQuality-R1-7B-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "reinforcement-learning", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2025-08-06", "format": "mlx", "mlx_only": true, "collection": "VisualQuality-R1", "description": "Image Quality Assessment", "_discovered": true }, { "name": "mlx-community/Holo1-3B-8bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2025-06-03", "format": "mlx", "mlx_only": true, "collection": "Holo1", "description": "", "_discovered": true }, { "name": "mlx-community/answerdotai-ModernBERT-Large-Instruct-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "fill-mask", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2025-04-02", "format": "mlx", "mlx_only": true, "collection": "ModernBert", "description": "", "_discovered": true }, { "name": "mlx-community/gemma-3-4b-pt-6bit", "provider": "mlx-community", "parameter_count": "4B", "parameters_raw": 4000000000, "min_ram_gb": 4.4, "recommended_ram_gb": 6.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2025-03-18", "format": "mlx", "mlx_only": true, "collection": "Gemma 3", "description": "A collection of lightweight, state-of-the-art open models built from the same research and technology that powers the Gemini 2.0 models", "_discovered": true }, { "name": "mlx-community/mamba2-2.7b-8bit", "provider": "mlx-community", "parameter_count": "2.7B", "parameters_raw": 2700000000, "min_ram_gb": 4.1, "recommended_ram_gb": 5.6, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2025-01-21", "format": "mlx", "mlx_only": true, "collection": "Mamba2", "description": "", "_discovered": true }, { "name": "mlx-community/EXAONE-3.5-2.4B-Instruct-8bit", "provider": "mlx-community", "parameter_count": "2.4B", "parameters_raw": 2400000000, "min_ram_gb": 3.8, "recommended_ram_gb": 5.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2024-12-09", "format": "mlx", "mlx_only": true, "collection": "EXAONE-3.5", "description": "EXAONE 3.5, a collection of instruction-tuned bilingual generative models ranging from 2.4B to 32B parameters, developed by LG AI.", "_discovered": true }, { "name": "mlx-community/Lumimaid-70B-v0.1-alt", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 41.2, "recommended_ram_gb": 49.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2024-10-13", "format": "mlx", "mlx_only": true, "collection": "Lumimaid", "description": "A collection of Neversleep's RP focused Lumimaid LLMs.", "_discovered": true }, { "name": "mlx-community/mamba-790m-hf-f32", "provider": "mlx-community", "parameter_count": "790M", "parameters_raw": 790000000, "min_ram_gb": 1.5, "recommended_ram_gb": 2.5, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 7, "hf_likes": 0, "release_date": "2024-09-21", "format": "mlx", "mlx_only": true, "collection": "Mamba", "description": "Mamba is a new LLM architecture that integrates the Structured State Space sequence model to manage lengthy data sequences.", "_discovered": true }, { "name": "mlx-community/functiongemma-270m-it-6bit", "provider": "mlx-community", "parameter_count": "270M", "parameters_raw": 270000000, "min_ram_gb": 1.2, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 2, "release_date": "2025-12-18", "format": "mlx", "mlx_only": true, "collection": "FunctionGemma", "description": "by Google Deepmind", "_discovered": true }, { "name": "mlx-community/QVQ-72B-Preview-6bit", "provider": "mlx-community", "parameter_count": "72B", "parameters_raw": 72000000000, "min_ram_gb": 63.1, "recommended_ram_gb": 74.9, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 6, "hf_likes": 2, "release_date": "2024-12-24", "format": "mlx", "mlx_only": true, "collection": "QVQ-72B-Preview", "description": "", "_discovered": true }, { "name": "mlx-community/EfRLFN-x2", "provider": "mlx-community", "parameter_count": "487.01K", "parameters_raw": 487010, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 6, "hf_likes": 1, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "EfRLFN MLX", "description": "MLX port of EfRLFN (ICLR 2026): realtime x2/x4 image super-resolution on Apple Silicon.", "_discovered": true }, { "name": "mlx-community/Llama-OuteTTS-1.0-1B-6bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.9, "recommended_ram_gb": 3.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 6, "hf_likes": 1, "release_date": "2025-05-19", "format": "mlx", "mlx_only": true, "collection": "OuteTTS-1.0", "description": "", "_discovered": true }, { "name": "mlx-community/reader-lm-1.5b", "provider": "mlx-community", "parameter_count": "1.5B", "parameters_raw": 1500000000, "min_ram_gb": 1.9, "recommended_ram_gb": 3.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 1, "release_date": "2025-01-18", "format": "mlx", "mlx_only": true, "collection": "Jina Reader-LM", "description": "Convert HTML content to LLM-friendly Markdown/JSON content", "_discovered": true }, { "name": "mlx-community/INTELLECT-3-8bit", "provider": "mlx-community", "parameter_count": "106.852B", "parameters_raw": 106852251264, "min_ram_gb": 123.9, "recommended_ram_gb": 146.3, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 0, "release_date": "2025-11-27", "format": "mlx", "mlx_only": true, "collection": "INTELLECT 3", "description": "", "_discovered": true }, { "name": "mlx-community/Olmo-3-7B-Instruct-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 0, "release_date": "2025-11-20", "format": "mlx", "mlx_only": true, "collection": "Olmo-3", "description": "Ai2's Olmo 3 model family of instruction and reasoning models.", "_discovered": true }, { "name": "mlx-community/Olmo-3-7B-Instruct-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 0, "release_date": "2025-11-20", "format": "mlx", "mlx_only": true, "collection": "Olmo-3", "description": "Ai2's Olmo 3 model family of instruction and reasoning models.", "_discovered": true }, { "name": "mlx-community/ERNIE-4.5-21B-A3B-PT-6bit", "provider": "mlx-community", "parameter_count": "21B", "parameters_raw": 21000000000, "min_ram_gb": 19.1, "recommended_ram_gb": 23.3, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 0, "release_date": "2025-07-04", "format": "mlx", "mlx_only": true, "collection": "ERNIE-4.5", "description": "", "_discovered": true }, { "name": "mlx-community/AceReason-Nemotron-7B-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 0, "release_date": "2025-05-26", "format": "mlx", "mlx_only": true, "collection": "AceReason Nemotron", "description": "", "_discovered": true }, { "name": "mlx-community/helium-1-preview-2b-float32", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 0, "release_date": "2025-01-18", "format": "mlx", "mlx_only": true, "collection": "Helium-1", "description": "Kyutai's Helium-1 2B Model, outperforming other state of the art small models.", "_discovered": true }, { "name": "mlx-community/encodec-24khz-bfloat16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 0, "release_date": "2024-09-18", "format": "mlx", "mlx_only": true, "collection": "EnCodec", "description": "EnCodec models in MLX", "_discovered": true }, { "name": "mlx-community/Yi-1.5-34B-8bit", "provider": "mlx-community", "parameter_count": "34B", "parameters_raw": 34000000000, "min_ram_gb": 40.1, "recommended_ram_gb": 47.9, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 6, "hf_likes": 0, "release_date": "2024-05-13", "format": "mlx", "mlx_only": true, "collection": "Yi-1.5", "description": "", "_discovered": true }, { "name": "mlx-community/P1-VL-30B-A3B-6bit", "provider": "mlx-community", "parameter_count": "30B", "parameters_raw": 30000000000, "min_ram_gb": 26.9, "recommended_ram_gb": 32.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2026-05-19", "format": "mlx", "mlx_only": true, "collection": "PRIME-RL P1-VL-30B-A3B", "description": "Bridging visual perception and scientific reasoning in physics olympiads", "_discovered": true }, { "name": "mlx-community/PE-Core-T16-384", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2025-12-26", "format": "mlx", "mlx_only": true, "collection": "Perception Encoder", "description": "Perception Encoder Models from Facebook", "_discovered": true }, { "name": "mlx-community/lille-130m-instruct-8bit", "provider": "mlx-community", "parameter_count": "130M", "parameters_raw": 130000000, "min_ram_gb": 1.1, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2025-09-05", "format": "mlx", "mlx_only": true, "collection": "Lille 130M", "description": "Very Small smart model created for the mobile", "_discovered": true }, { "name": "mlx-community/VisualQuality-R1-7B-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "reinforcement-learning", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2025-08-06", "format": "mlx", "mlx_only": true, "collection": "VisualQuality-R1", "description": "Image Quality Assessment", "_discovered": true }, { "name": "mlx-community/AceReason-Nemotron-7B-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2025-05-26", "format": "mlx", "mlx_only": true, "collection": "AceReason Nemotron", "description": "", "_discovered": true }, { "name": "mlx-community/answerdotai-ModernBERT-base-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "fill-mask", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2025-04-02", "format": "mlx", "mlx_only": true, "collection": "ModernBert", "description": "", "_discovered": true }, { "name": "mlx-community/UI-TARS-7B-SFT-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2025-03-03", "format": "mlx", "mlx_only": true, "collection": "UI-TARS", "description": "", "_discovered": true }, { "name": "mlx-community/UI-TARS-7B-SFT-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2025-03-03", "format": "mlx", "mlx_only": true, "collection": "UI-TARS", "description": "", "_discovered": true }, { "name": "mlx-community/reader-lm-0.5b", "provider": "mlx-community", "parameter_count": "500M", "parameters_raw": 500000000, "min_ram_gb": 1.3, "recommended_ram_gb": 2.3, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2025-01-18", "format": "mlx", "mlx_only": true, "collection": "Jina Reader-LM", "description": "Convert HTML content to LLM-friendly Markdown/JSON content", "_discovered": true }, { "name": "mlx-community/SmolVLM-Instruct-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2024-11-29", "format": "mlx", "mlx_only": true, "collection": "Idefics 3 + SmolVLM", "description": "", "_discovered": true }, { "name": "mlx-community/falcon-mamba-7b-8bit-instruct", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2024-11-15", "format": "mlx", "mlx_only": true, "collection": "Falcon-Mamba", "description": "Falcon Mamba models compatible with MLX", "_discovered": true }, { "name": "mlx-community/falcon-mamba-7b-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2024-11-15", "format": "mlx", "mlx_only": true, "collection": "Falcon-Mamba", "description": "Falcon Mamba models compatible with MLX", "_discovered": true }, { "name": "mlx-community/mamba-1.4b-hf-f32", "provider": "mlx-community", "parameter_count": "1.4B", "parameters_raw": 1400000000, "min_ram_gb": 1.8, "recommended_ram_gb": 2.9, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2024-09-21", "format": "mlx", "mlx_only": true, "collection": "Mamba", "description": "Mamba is a new LLM architecture that integrates the Structured State Space sequence model to manage lengthy data sequences.", "_discovered": true }, { "name": "mlx-community/encodec-32khz-bfloat16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "coding", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 5, "hf_likes": 0, "release_date": "2024-09-18", "format": "mlx", "mlx_only": true, "collection": "EnCodec", "description": "EnCodec models in MLX", "_discovered": true }, { "name": "mlx-community/plamo-2-8b-4bit", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4, "hf_likes": 3, "release_date": "2025-03-15", "format": "mlx", "mlx_only": true, "collection": "PLaMo", "description": "", "_discovered": true }, { "name": "mlx-community/OpenELM-1_1B-8bit", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4, "hf_likes": 1, "release_date": "2024-04-24", "format": "mlx", "mlx_only": true, "collection": "OpenELM", "description": "A family of Open-source Efficient Language Models from Apple.", "_discovered": true }, { "name": "mlx-community/Laguna-XS-2.1-8bit", "provider": "mlx-community", "parameter_count": "9.40587B", "parameters_raw": 9405869824, "min_ram_gb": 11.8, "recommended_ram_gb": 14.7, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4, "hf_likes": 0, "release_date": "2026-07-03", "format": "mlx", "mlx_only": true, "collection": "Laguna-XS-2.1", "description": "MLX versions of Laguna-XS-2.1", "_discovered": true }, { "name": "mlx-community/PE-Core-S16-384", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4, "hf_likes": 0, "release_date": "2025-12-26", "format": "mlx", "mlx_only": true, "collection": "Perception Encoder", "description": "Perception Encoder Models from Facebook", "_discovered": true }, { "name": "mlx-community/Apriel-1.5-15b-Thinker-5bit", "provider": "mlx-community", "parameter_count": "15B", "parameters_raw": 15000000000, "min_ram_gb": 11.8, "recommended_ram_gb": 14.7, "min_vram_gb": 0.0, "quantization": "mlx-5bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4, "hf_likes": 0, "release_date": "2025-10-03", "format": "mlx", "mlx_only": true, "collection": "ServiceNow-Apriel", "description": "Apriel-1.5-15b-Thinker is a multimodal reasoning model in ServiceNow’s Apriel SLM series which achieves competitive performance against models 10 time", "_discovered": true }, { "name": "mlx-community/lille-130m-instruct-6bit", "provider": "mlx-community", "parameter_count": "130M", "parameters_raw": 130000000, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4, "hf_likes": 0, "release_date": "2025-09-05", "format": "mlx", "mlx_only": true, "collection": "Lille 130M", "description": "Very Small smart model created for the mobile", "_discovered": true }, { "name": "mlx-community/VisualQuality-R1-7B-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "reasoning", "capabilities": [ "mlx" ], "pipeline_tag": "reinforcement-learning", "architecture": "", "hf_downloads": 4, "hf_likes": 0, "release_date": "2025-08-06", "format": "mlx", "mlx_only": true, "collection": "VisualQuality-R1", "description": "Image Quality Assessment", "_discovered": true }, { "name": "mlx-community/Virtuoso-Medium-v2-3bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 4.0, "recommended_ram_gb": 5.5, "min_vram_gb": 0.0, "quantization": "mlx-3bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4, "hf_likes": 0, "release_date": "2025-01-30", "format": "mlx", "mlx_only": true, "collection": "Arcee Virtuoso", "description": "", "_discovered": true }, { "name": "mlx-community/helium-1-preview-2b", "provider": "mlx-community", "parameter_count": "2B", "parameters_raw": 2000000000, "min_ram_gb": 2.1, "recommended_ram_gb": 3.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 4, "hf_likes": 0, "release_date": "2025-01-18", "format": "mlx", "mlx_only": true, "collection": "Helium-1", "description": "Kyutai's Helium-1 2B Model, outperforming other state of the art small models.", "_discovered": true }, { "name": "mlx-community/Perception-LM-3B", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 2.7, "recommended_ram_gb": 4.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 3, "hf_likes": 0, "release_date": "2026-01-07", "format": "mlx", "mlx_only": true, "collection": "facebook Perception LM", "description": "A collection of facebook perception language models", "_discovered": true }, { "name": "mlx-community/PE-Core-L14-336", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 3, "hf_likes": 0, "release_date": "2025-12-26", "format": "mlx", "mlx_only": true, "collection": "Perception Encoder", "description": "Perception Encoder Models from Facebook", "_discovered": true }, { "name": "mlx-community/EXAONE-3.5-2.4B-Instruct-6bit", "provider": "mlx-community", "parameter_count": "2.4B", "parameters_raw": 2400000000, "min_ram_gb": 3.1, "recommended_ram_gb": 4.4, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "chat", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 3, "hf_likes": 0, "release_date": "2024-12-09", "format": "mlx", "mlx_only": true, "collection": "EXAONE-3.5", "description": "EXAONE 3.5, a collection of instruction-tuned bilingual generative models ranging from 2.4B to 32B parameters, developed by LG AI.", "_discovered": true }, { "name": "mlx-community/paligemma2-3b-ft-docci-448-6bit", "provider": "mlx-community", "parameter_count": "3B", "parameters_raw": 3000000000, "min_ram_gb": 3.6, "recommended_ram_gb": 5.0, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "image-text-to-text", "architecture": "", "hf_downloads": 3, "hf_likes": 0, "release_date": "2024-12-05", "format": "mlx", "mlx_only": true, "collection": "Paligemma 2", "description": "", "_discovered": true }, { "name": "mlx-community/plamo-2-8b", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2, "hf_likes": 2, "release_date": "2025-03-16", "format": "mlx", "mlx_only": true, "collection": "PLaMo", "description": "", "_discovered": true }, { "name": "mlx-community/Perception-LM-1B", "provider": "mlx-community", "parameter_count": "1B", "parameters_raw": 1000000000, "min_ram_gb": 1.6, "recommended_ram_gb": 2.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2, "hf_likes": 0, "release_date": "2026-01-07", "format": "mlx", "mlx_only": true, "collection": "facebook Perception LM", "description": "A collection of facebook perception language models", "_discovered": true }, { "name": "mlx-community/Virtuoso-Medium-v2-6bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 7.0, "recommended_ram_gb": 9.1, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 2, "hf_likes": 0, "release_date": "2025-01-30", "format": "mlx", "mlx_only": true, "collection": "Arcee Virtuoso", "description": "", "_discovered": true }, { "name": "mlx-community/Lumimaid-70B-v0.1", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 41.2, "recommended_ram_gb": 49.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1, "hf_likes": 0, "release_date": "2024-10-13", "format": "mlx", "mlx_only": true, "collection": "Lumimaid", "description": "A collection of Neversleep's RP focused Lumimaid LLMs.", "_discovered": true }, { "name": "mlx-community/Lumimaid-70B-v0.1-OAS", "provider": "mlx-community", "parameter_count": "70B", "parameters_raw": 70000000000, "min_ram_gb": 41.2, "recommended_ram_gb": 49.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 1, "hf_likes": 0, "release_date": "2024-10-13", "format": "mlx", "mlx_only": true, "collection": "Lumimaid", "description": "A collection of Neversleep's RP focused Lumimaid LLMs.", "_discovered": true }, { "name": "mlx-community/demucs-mlx-fp16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 0, "hf_likes": 9, "release_date": "2026-03-16", "format": "mlx", "mlx_only": true, "collection": "Demucs MLX — Music Source Separation", "description": "Demucs music stem separation for Apple Silicon. Float32 and float16 variants.", "_discovered": true }, { "name": "mlx-community/demucs-mlx", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "audio-to-audio", "architecture": "", "hf_downloads": 0, "hf_likes": 7, "release_date": "2026-03-16", "format": "mlx", "mlx_only": true, "collection": "Demucs MLX — Music Source Separation", "description": "Demucs music stem separation for Apple Silicon. Float32 and float16 variants.", "_discovered": true }, { "name": "mlx-community/Boogu-Image-0.1-Base-4bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 5, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Boogu-Image-0.1 (MLX)", "description": "MLX conversions of Boogu-Image-0.1 (OmniGen2-lineage T2I/edit, Apache-2.0) for Apple Silicon.", "_discovered": true }, { "name": "mlx-community/LongCat-Video-Avatar-1.5-bf16-dmd-merged", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-video", "architecture": "", "hf_downloads": 0, "hf_likes": 5, "release_date": "2026-05-28", "format": "mlx", "mlx_only": true, "collection": "LongCat-Video-Avatar 1.5 — MLX", "description": "Apple MLX port of Meituan's audio-driven video diffusion. Source + recipe: github.com/xocialize/longcat-avatar-mlx", "_discovered": true }, { "name": "mlx-community/supertonic-3", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "tts", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-speech", "architecture": "", "hf_downloads": 0, "hf_likes": 4, "release_date": "2026-05-30", "format": "mlx", "mlx_only": true, "collection": "Supertonic 3", "description": "by Supertone, converted to MLX", "_discovered": true }, { "name": "mlx-community/Wan2.2-VAE-Lance-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 3, "release_date": "2026-05-21", "format": "mlx", "mlx_only": true, "collection": "Lance MLX", "description": "Feature-complete MLX port of ByteDance Lance: t2i, image_edit, x2t_image, t2v, video_edit, x2t_video.", "_discovered": true }, { "name": "mlx-community/Boogu-Image-0.1-Base-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 2, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Boogu-Image-0.1 (MLX)", "description": "MLX conversions of Boogu-Image-0.1 (OmniGen2-lineage T2I/edit, Apache-2.0) for Apple Silicon.", "_discovered": true }, { "name": "mlx-community/LongCat-Video-q8", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-video", "architecture": "", "hf_downloads": 0, "hf_likes": 2, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "LongCat-Video — MLX", "description": "Apple MLX port of Meituan's 13.6B base text-to-video model. Six task variants share one DiT. github.com/xocialize/longcat-video-mlx", "_discovered": true }, { "name": "mlx-community/Boogu-Image-0.1-Turbo-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 1, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Boogu-Image-0.1 (MLX)", "description": "MLX conversions of Boogu-Image-0.1 (OmniGen2-lineage T2I/edit, Apache-2.0) for Apple Silicon.", "_discovered": true }, { "name": "mlx-community/Boogu-Image-0.1-Turbo-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 1, "release_date": "2026-06-21", "format": "mlx", "mlx_only": true, "collection": "Boogu-Image-0.1 (MLX)", "description": "MLX conversions of Boogu-Image-0.1 (OmniGen2-lineage T2I/edit, Apache-2.0) for Apple Silicon.", "_discovered": true }, { "name": "mlx-community/LFM2-VL-450M-6bit", "provider": "mlx-community", "parameter_count": "450M", "parameters_raw": 450000000, "min_ram_gb": 1.4, "recommended_ram_gb": 2.5, "min_vram_gb": 0.0, "quantization": "mlx-6bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 0, "hf_likes": 1, "release_date": "2025-08-16", "format": "mlx", "mlx_only": true, "collection": "LFM2-VL", "description": "", "_discovered": true }, { "name": "mlx-community/LongCat-Video-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-video", "architecture": "", "hf_downloads": 0, "hf_likes": 1, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "LongCat-Video — MLX", "description": "Apple MLX port of Meituan's 13.6B base text-to-video model. Six task variants share one DiT. github.com/xocialize/longcat-video-mlx", "_discovered": true }, { "name": "mlx-community/LongCat-Video-q4", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-video", "architecture": "", "hf_downloads": 0, "hf_likes": 1, "release_date": "2026-06-04", "format": "mlx", "mlx_only": true, "collection": "LongCat-Video — MLX", "description": "Apple MLX port of Meituan's 13.6B base text-to-video model. Six task variants share one DiT. github.com/xocialize/longcat-video-mlx", "_discovered": true }, { "name": "mlx-community/whisper-large-v2-mlx-fp32", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "stt", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 0, "hf_likes": 1, "release_date": "2024-08-09", "format": "mlx", "mlx_only": true, "collection": "Whisper", "description": "OpenAI Whisper speech recognition models in MLX format", "_discovered": true }, { "name": "mlx-community/MI-GAN-512-places2-fp16", "provider": "mlx-community", "parameter_count": "7.37137M", "parameters_raw": 7371368, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-06-24", "format": "mlx", "mlx_only": true, "collection": "Inpainting (MLX)", "description": "Apple-MLX fp16 inpainting / object-removal models (LaMa Apache-2.0 + MI-GAN MIT). Loaded by mlx-lama-swift.", "_discovered": true }, { "name": "mlx-community/MI-GAN-256-places2-fp16", "provider": "mlx-community", "parameter_count": "6.29305M", "parameters_raw": 6293045, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-06-24", "format": "mlx", "mlx_only": true, "collection": "Inpainting (MLX)", "description": "Apple-MLX fp16 inpainting / object-removal models (LaMa Apache-2.0 + MI-GAN MIT). Loaded by mlx-lama-swift.", "_discovered": true }, { "name": "mlx-community/MI-GAN-256-ffhq-fp16", "provider": "mlx-community", "parameter_count": "6.29305M", "parameters_raw": 6293045, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-06-24", "format": "mlx", "mlx_only": true, "collection": "Inpainting (MLX)", "description": "Apple-MLX fp16 inpainting / object-removal models (LaMa Apache-2.0 + MI-GAN MIT). Loaded by mlx-lama-swift.", "_discovered": true }, { "name": "mlx-community/LaMa-bf16", "provider": "mlx-community", "parameter_count": "51.057M", "parameters_raw": 51057027, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-06-24", "format": "mlx", "mlx_only": true, "collection": "Inpainting (MLX)", "description": "Apple-MLX fp16 inpainting / object-removal models (LaMa Apache-2.0 + MI-GAN MIT). Loaded by mlx-lama-swift.", "_discovered": true }, { "name": "mlx-community/DDColor-modelscope-fp16", "provider": "mlx-community", "parameter_count": "227.882M", "parameters_raw": 227881750, "min_ram_gb": 1.1, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-06-24", "format": "mlx", "mlx_only": true, "collection": "DDColor (MLX)", "description": "Apple-MLX fp16 builds of DDColor automatic image colorization (piddnad/DDColor, Apache-2.0). Loaded by mlx-ddcolor-swift.", "_discovered": true }, { "name": "mlx-community/DDColor-paper-tiny-fp16", "provider": "mlx-community", "parameter_count": "55.0193M", "parameters_raw": 55019254, "min_ram_gb": 1.0, "recommended_ram_gb": 2.0, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-06-24", "format": "mlx", "mlx_only": true, "collection": "DDColor (MLX)", "description": "Apple-MLX fp16 builds of DDColor automatic image colorization (piddnad/DDColor, Apache-2.0). Loaded by mlx-ddcolor-swift.", "_discovered": true }, { "name": "mlx-community/DDColor-artistic-fp16", "provider": "mlx-community", "parameter_count": "227.882M", "parameters_raw": 227881750, "min_ram_gb": 1.1, "recommended_ram_gb": 2.2, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-to-image", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-06-24", "format": "mlx", "mlx_only": true, "collection": "DDColor (MLX)", "description": "Apple-MLX fp16 builds of DDColor automatic image colorization (piddnad/DDColor, Apache-2.0). Loaded by mlx-ddcolor-swift.", "_discovered": true }, { "name": "mlx-community/BiRefNet-fp16", "provider": "mlx-community", "parameter_count": "220.203M", "parameters_raw": 220202578, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-segmentation", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-06-22", "format": "mlx", "mlx_only": true, "collection": "BiRefNet (MLX)", "description": "fp16 MLX BiRefNet matting: general @1024 (fast) + HR-matting @2048 (best). MIT. Loaded by xocialize/mlx-birefnet-swift.", "_discovered": true }, { "name": "mlx-community/BiRefNet_HR-matting-fp16", "provider": "mlx-community", "parameter_count": "220.203M", "parameters_raw": 220202578, "min_ram_gb": 1.1, "recommended_ram_gb": 2.1, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-segmentation", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-06-22", "format": "mlx", "mlx_only": true, "collection": "BiRefNet (MLX)", "description": "fp16 MLX BiRefNet matting: general @1024 (fast) + HR-matting @2048 (best). MIT. Loaded by xocialize/mlx-birefnet-swift.", "_discovered": true }, { "name": "mlx-community/LongCat-Video-Avatar-1.5-bf16", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 17.1, "recommended_ram_gb": 20.9, "min_vram_gb": 0.0, "quantization": "BF16", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-to-video", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-05-28", "format": "mlx", "mlx_only": true, "collection": "LongCat-Video-Avatar 1.5 — MLX", "description": "Apple MLX port of Meituan's audio-driven video diffusion. Source + recipe: github.com/xocialize/longcat-avatar-mlx", "_discovered": true }, { "name": "mlx-community/Perception-LM-8B", "provider": "mlx-community", "parameter_count": "8B", "parameters_raw": 8000000000, "min_ram_gb": 5.6, "recommended_ram_gb": 7.4, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2026-01-07", "format": "mlx", "mlx_only": true, "collection": "facebook Perception LM", "description": "A collection of facebook perception language models", "_discovered": true }, { "name": "mlx-community/simclrv1-imagenet1k-resnet50-1x", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-classification", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2025-05-14", "format": "mlx", "mlx_only": true, "collection": "SimCLRv1", "description": "", "_discovered": true }, { "name": "mlx-community/simclrv1-imagenet1k-resnet50-2x", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-classification", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2025-05-14", "format": "mlx", "mlx_only": true, "collection": "SimCLRv1", "description": "", "_discovered": true }, { "name": "mlx-community/simclrv1-imagenet1k-resnet50-4x", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 5.0, "recommended_ram_gb": 6.7, "min_vram_gb": 0.0, "quantization": "mlx-4bit", "context_length": 32768, "use_case": "general", "capabilities": [ "mlx" ], "pipeline_tag": "image-classification", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2025-05-14", "format": "mlx", "mlx_only": true, "collection": "SimCLRv1", "description": "", "_discovered": true }, { "name": "mlx-community/olmOCR-7B-0225-preview-8bit", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2025-03-03", "format": "mlx", "mlx_only": true, "collection": "olmOCR", "description": "", "_discovered": true }, { "name": "mlx-community/Molmo-7B-D-0924-8bit-skip-vision", "provider": "mlx-community", "parameter_count": "7B", "parameters_raw": 7000000000, "min_ram_gb": 9.0, "recommended_ram_gb": 11.5, "min_vram_gb": 0.0, "quantization": "mlx-8bit", "context_length": 32768, "use_case": "multimodal", "capabilities": [ "mlx" ], "pipeline_tag": "text-generation", "architecture": "", "hf_downloads": 0, "hf_likes": 0, "release_date": "2024-11-20", "format": "mlx", "mlx_only": true, "collection": "Molmo", "description": "", "_discovered": true } ]