{
  "slug": "mistral-large-3",
  "order": 28,
  "name": "Mistral Large 3 (673B)",
  "baseTitle": "Mistral Large 3",
  "company": "Mistral AI",
  "country": "France",
  "flag": "🇫🇷",
  "date": "2025-12-02",
  "decoderType": "Sparse MoE",
  "attention": "MLA",
  "layerMix": "61 MLA",
  "scale": "673B total, 41B active (6.1% active)",
  "totalParams": 673000000000,
  "activeParams": 41000000000,
  "context": "262,144",
  "contextTokens": 262144,
  "license": "Apache License 2.0",
  "kvCacheText": "68.6 KiB",
  "kvCacheKiB": 68.6,
  "kvCacheLevel": "Low",
  "summary": "Mistral's new flagship effectively adopts the DeepSeek architecture and retunes the expert sizes.",
  "keyDetail": "Near-clone of DeepSeek V3 with larger experts, fewer routed experts, and multimodal support.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/mistral-3-large-673-billion.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/mistral-3-large-673-billion.webp",
    "alt": "Figure 50: Mistral Large 3",
    "width": 960,
    "height": 942
  },
  "articleUrl": "https://magazine.sebastianraschka.com/p/the-big-llm-architecture-comparison#%C2%A717-mistral-3-large",
  "links": [
    {
      "label": "View in article",
      "href": "https://magazine.sebastianraschka.com/p/the-big-llm-architecture-comparison#%C2%A717-mistral-3-large",
      "title": null
    },
    {
      "label": "params.json",
      "href": "https://huggingface.co/mistralai/Mistral-Large-3-675B-Instruct-2512/blob/main/params.json",
      "title": "mistralai/Mistral-Large-3-675B-Instruct-2512"
    }
  ],
  "concepts": [
    {
      "slug": "mla",
      "label": "MLA",
      "name": "Multi-Head Latent Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/mla/"
    },
    {
      "slug": "moe",
      "label": "MoE",
      "name": "Mixture of Experts",
      "url": "https://llm-atlas-7or.pages.dev/concepts/moe/"
    }
  ],
  "aai": {
    "v4": {
      "label": "AA Intelligence Index v4.3.2",
      "total": 9.3,
      "profile": {
        "General": 17.1,
        "Scientific": 2.1,
        "Coding": 18.3,
        "Agents": 0.2
      },
      "url": "https://artificialanalysis.ai/models/mistral-large-3"
    }
  },
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-mistral-large-3",
  "family": "moe",
  "url": "https://llm-atlas-7or.pages.dev/models/mistral-large-3/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/mistral-large-3.md"
}