{
  "version": 1,
  "updated": "2026-10-19",
  "models": [
    {
      "id": "mlx-community/gemma-3-1b-it-qat-4bit",
      "name": "Gemma 3 · 1B",
      "icon": "gemma",
      "sizeGB": 0.8,
      "architecture": "gemma3_text",
      "free": true,
      "proFrom": "1.2",
      "quality": 30,
      "extraEOSTokens": [
        "<end_of_turn>"
      ],
      "strengths": [
        "Chatting and everyday questions",
        "Writing and summarizing",
        "Translating 140+ languages"
      ]
    },
    {
      "id": "mlx-community/Llama-3.2-1B-Instruct-4bit",
      "name": "Llama 3.2 · 1B",
      "icon": "llama",
      "sizeGB": 0.7,
      "architecture": "llama",
      "freeFrom": "1.2",
      "quality": 20,
      "strengths": [
        "Quick chats and rewriting",
        "Summarizing short texts",
        "Fast on older iPhones"
      ]
    },
    {
      "id": "mlx-community/Llama-3.2-3B-Instruct-4bit",
      "name": "Llama 3.2 · 3B",
      "icon": "llama",
      "sizeGB": 1.8,
      "architecture": "llama",
      "quality": 45,
      "strengths": [
        "Chatting and writing",
        "Summarizing and following instructions",
        "Better answers than the 1B"
      ]
    },
    {
      "id": "mlx-community/Qwen3-0.6B-4bit",
      "name": "Qwen 3 · 0.6B",
      "icon": "qwen",
      "sizeGB": 0.4,
      "architecture": "qwen3",
      "reasoning": true,
      "quality": 15,
      "strengths": [
        "Very fast, tiny size",
        "Simple questions and rewriting",
        "Can show its reasoning"
      ]
    },
    {
      "id": "mlx-community/Qwen3-1.7B-4bit",
      "name": "Qwen 3 · 1.7B",
      "icon": "qwen",
      "sizeGB": 1.0,
      "architecture": "qwen3",
      "reasoning": true,
      "quality": 35,
      "strengths": [
        "Reasoning step by step",
        "Chatting and summarizing",
        "Strong in Chinese and 100+ languages"
      ]
    },
    {
      "id": "mlx-community/Qwen3-4B-4bit",
      "name": "Qwen 3 · 4B",
      "icon": "qwen",
      "sizeGB": 2.3,
      "architecture": "qwen3",
      "reasoning": true,
      "quality": 60,
      "strengths": [
        "Reasoning, maths and code",
        "Writing and summarizing",
        "Translating 100+ languages"
      ]
    },
    {
      "id": "mlx-community/Qwen3-8B-4bit",
      "name": "Qwen 3 · 8B",
      "icon": "qwen",
      "sizeGB": 4.7,
      "architecture": "qwen3",
      "reasoning": true,
      "quality": 75,
      "strengths": [
        "Hard reasoning, maths and code",
        "Long, detailed answers",
        "Translating 100+ languages"
      ]
    },
    {
      "id": "mlx-community/LFM2.5-1.2B-Instruct-4bit",
      "name": "LFM2.5 · 1.2B",
      "icon": "liquid",
      "sizeGB": 0.7,
      "architecture": "lfm2",
      "quality": 28,
      "strengths": [
        "Fast, efficient chat",
        "Summarizing and extracting info",
        "Low battery use"
      ]
    },
    {
      "id": "mlx-community/LFM2-1.2B-4bit",
      "name": "LFM2 · 1.2B",
      "icon": "liquid",
      "sizeGB": 0.7,
      "architecture": "lfm2",
      "quality": 25,
      "strengths": [
        "Fast, efficient chat",
        "Summarizing and extracting info",
        "Low battery use"
      ],
      "status": "hidden"
    },
    {
      "id": "mlx-community/SmolLM3-3B-4bit",
      "name": "SmolLM3 · 3B",
      "icon": "smollm",
      "sizeGB": 1.8,
      "architecture": "smollm3",
      "reasoning": true,
      "quality": 40,
      "strengths": [
        "Reasoning and long texts",
        "Chatting in 6 languages",
        "Summarizing"
      ]
    },
    {
      "id": "mlx-community/gemma-3n-E2B-it-lm-4bit",
      "name": "Gemma 3n · E2B",
      "icon": "gemma",
      "sizeGB": 2.6,
      "architecture": "gemma3n",
      "quality": 55,
      "extraEOSTokens": [
        "<end_of_turn>"
      ],
      "strengths": [
        "Chatting and writing",
        "Translating 140+ languages",
        "Quality of a bigger model, light on memory"
      ]
    },
    {
      "id": "mlx-community/gemma-4-e2b-it-4bit",
      "name": "Gemma 4 · E2B",
      "icon": "gemma",
      "sizeGB": 3.6,
      "architecture": "gemma4",
      "quality": 70,
      "strengths": [
        "Chatting and writing",
        "Writing and summarizing",
        "Translating 140+ languages"
      ]
    },
    {
      "id": "mlx-community/DeepSeek-R1-Distill-Qwen-1.5B-4bit",
      "name": "DeepSeek R1 Distill 1.5B",
      "icon": "deepseek",
      "sizeGB": 1.0,
      "architecture": "qwen2",
      "reasoning": true,
      "experimental": true,
      "quality": 10,
      "strengths": [
        "Reasoning model. May ramble on small devices."
      ],
      "status": "hidden"
    }
  ]
}