{
  "slug": "deepseek-v3",
  "order": 4,
  "name": "DeepSeek V3 (671B)",
  "baseTitle": "DeepSeek V3",
  "company": "DeepSeek",
  "country": "China",
  "flag": "🇨🇳",
  "date": "2024-12-26",
  "decoderType": "Sparse MoE",
  "attention": "MLA",
  "layerMix": "61 MLA",
  "scale": "671B total, 37B active (5.5% active)",
  "totalParams": 671000000000,
  "activeParams": 37000000000,
  "context": "128,000",
  "contextTokens": 128000,
  "license": "DeepSeek License Agreement v1.0",
  "kvCacheText": "68.6 KiB",
  "kvCacheKiB": 68.6,
  "kvCacheLevel": "Low",
  "summary": "DeepSeek's flagship template kicked off the recent wave of large open MoE models.",
  "keyDetail": "Uses a dense prefix, shared expert, and MTP-1 objective to keep a very large model practical.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/deepseek-v3-r1-671-billion.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/deepseek-v3-r1-671-billion.webp",
    "alt": "Figure 17: DeepSeek V3",
    "width": 960,
    "height": 938
  },
  "articleUrl": "https://magazine.sebastianraschka.com/p/the-big-llm-architecture-comparison#%C2%A75-llama-4",
  "links": [
    {
      "label": "View in article",
      "href": "https://magazine.sebastianraschka.com/p/the-big-llm-architecture-comparison#%C2%A75-llama-4",
      "title": null
    },
    {
      "label": "config.json",
      "href": "https://huggingface.co/deepseek-ai/DeepSeek-V3/blob/main/config.json",
      "title": "deepseek-ai/DeepSeek-V3"
    },
    {
      "label": "License",
      "href": "https://huggingface.co/deepseek-ai/DeepSeek-V3/blame/main/LICENSE-MODEL",
      "title": null
    },
    {
      "label": "Tech report",
      "href": "https://arxiv.org/pdf/2412.19437",
      "title": null
    }
  ],
  "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/"
    },
    {
      "slug": "mtp",
      "label": "MTP",
      "name": "Multi-Token Prediction",
      "url": "https://llm-atlas-7or.pages.dev/concepts/mtp/"
    }
  ],
  "aai": {
    "v4": {
      "label": "AA Intelligence Index v4.3.2",
      "total": 8.5,
      "profile": {
        "General": 15.3,
        "Scientific": 1.5,
        "Coding": 17.9,
        "Agents": 0.1
      },
      "url": "https://artificialanalysis.ai/models/deepseek-v3"
    }
  },
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-deepseek-v3",
  "family": "moe",
  "url": "https://llm-atlas-7or.pages.dev/models/deepseek-v3/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/deepseek-v3.md"
}