{
  "slug": "inkling",
  "order": 86,
  "name": "Inkling (975B)",
  "baseTitle": "Inkling",
  "company": "Thinking Machines Lab",
  "country": "United States",
  "flag": "🇺🇸",
  "date": "2026-07-15",
  "decoderType": "Sparse MoE",
  "attention": "GQA with QK-Norm and learned relative-position bias in a 5:1 sliding-window/global pattern",
  "layerMix": "55 sliding-window + 11 global GQA",
  "scale": "975B total, 41B active (4.2% active)",
  "totalParams": 975000000000,
  "activeParams": 41000000000,
  "context": "1,048,576",
  "contextTokens": 1048576,
  "license": "Apache License 2.0",
  "kvCacheText": "484 KiB",
  "kvCacheKiB": 484,
  "kvCacheLevel": "Very high",
  "summary": "Thinking Machines Lab's native multimodal model combines a 66-layer sparse MoE decoder with a 5:1 mix of local and global GQA.",
  "keyDetail": "Uses learned relative-position bias instead of RoPE and adds four short convolutions to each decoder layer.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/inkling.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/inkling.webp",
    "alt": "Figure 1: Inkling",
    "width": 930,
    "height": 960
  },
  "articleUrl": null,
  "links": [
    {
      "label": "config.json",
      "href": "https://huggingface.co/thinkingmachines/Inkling/blob/main/config.json",
      "title": "thinkingmachines/Inkling"
    },
    {
      "label": "License",
      "href": "https://www.apache.org/licenses/LICENSE-2.0",
      "title": null
    },
    {
      "label": "Tech report",
      "href": "https://thinkingmachines.ai/news/introducing-inkling/",
      "title": null
    }
  ],
  "concepts": [
    {
      "slug": "qk-norm",
      "label": "QK-Norm",
      "name": "QK-Norm",
      "url": "https://llm-atlas-7or.pages.dev/concepts/qk-norm/"
    },
    {
      "slug": "gqa",
      "label": "GQA",
      "name": "Grouped-Query Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/gqa/"
    },
    {
      "slug": "swa",
      "label": "SWA",
      "name": "Sliding Window Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/swa/"
    },
    {
      "slug": "shortconv",
      "label": "ShortConv",
      "name": "Short Convolution",
      "url": "https://llm-atlas-7or.pages.dev/concepts/shortconv/"
    },
    {
      "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": {
    "v5": {
      "label": "AA Intelligence Index v5.0",
      "total": 12.5,
      "profile": {
        "Knowledge Work": 13.5,
        "Coding": 2.8,
        "Science": 7.2,
        "General": 26.5
      },
      "url": "https://artificialanalysis.ai/models/inkling"
    },
    "v4": {
      "label": "AA Intelligence Index v4.3.2",
      "total": 25,
      "profile": {
        "General": 36.4,
        "Scientific": 18.7,
        "Coding": 24,
        "Agents": 18.7
      },
      "url": "https://artificialanalysis.ai/models/inkling"
    }
  },
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-inkling",
  "family": "moe",
  "url": "https://llm-atlas-7or.pages.dev/models/inkling/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/inkling.md"
}