{
  "slug": "qwen3-8-27b",
  "order": 97,
  "name": "Qwen3.8 (27B)",
  "baseTitle": "Qwen3.8 27B",
  "company": "Alibaba Cloud",
  "country": "China",
  "flag": "🇨🇳",
  "date": "2026-08-14",
  "decoderType": "Dense hybrid",
  "attention": "3:1 Gated DeltaNet and Gated Attention",
  "layerMix": "16 gated attention + 48 DeltaNet",
  "scale": "27B parameters",
  "totalParams": 27000000000,
  "activeParams": 27000000000,
  "context": "262,144",
  "contextTokens": 262144,
  "license": "Apache License 2.0",
  "kvCacheText": "64 KiB",
  "kvCacheKiB": 64,
  "kvCacheLevel": "Low",
  "summary": "Dense 27B native vision-language model with a 64-layer Gated DeltaNet and Gated Attention language backbone; the figure omits the vision encoder and MTP head.",
  "keyDetail": "Uses a 64-layer 3:1 Gated DeltaNet/Gated Attention decoder with partial RoPE and one MTP layer.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/qwen3-8-27b.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/qwen3-8-27b.webp",
    "alt": "Figure 1: Qwen3.8 27B",
    "width": 960,
    "height": 915
  },
  "articleUrl": null,
  "links": [
    {
      "label": "config.json",
      "href": "https://huggingface.co/Qwen/Qwen3.8-27B/blob/main/config.json",
      "title": "Qwen/Qwen3.8-27B"
    },
    {
      "label": "License",
      "href": "https://huggingface.co/Qwen/Qwen3.8-27B/blob/main/LICENSE",
      "title": null
    }
  ],
  "concepts": [
    {
      "slug": "hybrid-attention",
      "label": "Hybrid Attention",
      "name": "Hybrid Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/hybrid-attention/"
    },
    {
      "slug": "gqa",
      "label": "GQA",
      "name": "Grouped-Query Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/gqa/"
    },
    {
      "slug": "gated-attention",
      "label": "Gated Attention",
      "name": "Gated Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/gated-attention/"
    },
    {
      "slug": "qk-norm",
      "label": "QK-Norm",
      "name": "QK-Norm",
      "url": "https://llm-atlas-7or.pages.dev/concepts/qk-norm/"
    },
    {
      "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": 20.1,
      "profile": {
        "Knowledge Work": 39.7,
        "Coding": 11.2,
        "Science": 1.9,
        "General": 27.3
      },
      "url": "https://artificialanalysis.ai/models/qwen3-8-27b"
    },
    "v4": {
      "label": "AA Intelligence Index v4.3.2",
      "total": 33.7,
      "profile": {
        "General": 35.9,
        "Scientific": 19.7,
        "Coding": 26.1,
        "Agents": 45.8
      },
      "url": "https://artificialanalysis.ai/models/qwen3-8-27b"
    }
  },
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-qwen3-8-27b",
  "family": "hybrid",
  "url": "https://llm-atlas-7or.pages.dev/models/qwen3-8-27b/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/qwen3-8-27b.md"
}