{
  "slug": "gemma-4-12b",
  "order": 37,
  "name": "Gemma 4 (12B)",
  "baseTitle": "Gemma 4 12B",
  "company": "Google",
  "country": "United States",
  "flag": "🇺🇸",
  "date": "2026-06-03",
  "decoderType": "Dense",
  "attention": "GQA with QK-Norm, unified K/V on global layers, p-RoPE on global layers, and 5:1 sliding-window/global attention",
  "layerMix": "40 sliding-window + 8 global",
  "scale": "12B parameters",
  "totalParams": 12000000000,
  "activeParams": 12000000000,
  "context": "128,000",
  "contextTokens": 128000,
  "license": "Apache License 2.0",
  "kvCacheText": "328 KiB",
  "kvCacheKiB": 328,
  "kvCacheLevel": "Very high",
  "summary": "Dense Gemma 4 12B Unified uses an encoder-free multimodal input path, while its text decoder keeps the 31B-style local-global attention recipe.",
  "keyDetail": "Figure shows the 48-layer text decoder; the full 12B Unified model also projects audio, image, and video inputs directly into the decoder input stream.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/gemma-4-12b.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/gemma-4-12b.webp",
    "alt": "Figure 59: Gemma 4 12B",
    "width": 918,
    "height": 960
  },
  "articleUrl": null,
  "links": [
    {
      "label": "config.json",
      "href": "https://huggingface.co/google/gemma-4-12B-it/blob/main/config.json",
      "title": "google/gemma-4-12B-it"
    },
    {
      "label": "Tech report",
      "href": "https://developers.googleblog.com/gemma-4-12b-the-developer-guide/",
      "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/"
    }
  ],
  "aai": {
    "v4": {
      "label": "AA Intelligence Index v4.3.2",
      "total": 9.4,
      "profile": {
        "Scientific": 3.2
      },
      "url": "https://artificialanalysis.ai/models/gemma-4-12b-non-reasoning"
    }
  },
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-gemma-4-12b",
  "family": "dense",
  "url": "https://llm-atlas-7or.pages.dev/models/gemma-4-12b/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/gemma-4-12b.md"
}