{
  "slug": "tencent-hy3-preview-295b-a21b",
  "order": 71,
  "name": "Tencent Hy3-preview (295B-A21B)",
  "baseTitle": "Tencent Hy3-preview 295B-A21B",
  "company": "Tencent",
  "country": "China",
  "flag": "🇨🇳",
  "date": "2026-04-23",
  "decoderType": "Sparse MoE",
  "attention": "GQA with QK-Norm",
  "layerMix": "80 GQA + 1 MTP layer",
  "scale": "295B total, 21B active (7.1% active)",
  "totalParams": 295000000000,
  "activeParams": 21000000000,
  "context": "262,144",
  "contextTokens": 262144,
  "license": "Tencent Hy Community License Agreement",
  "kvCacheText": "320 KiB",
  "kvCacheKiB": 320,
  "kvCacheLevel": "Very high",
  "summary": "Tencent Hunyuan's 295B sparse MoE preview combines a dense first FFN layer, GQA with QK-Norm, and one extra MTP layer for faster generation.",
  "keyDetail": "Uses 192 routed experts, 8 routed plus 1 shared expert active per token, sigmoid routing with expert bias, and a 3.8B MTP layer.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/tencent-hy3.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/tencent-hy3.webp",
    "alt": "Figure 1: Tencent Hy3-preview 295B-A21B",
    "width": 960,
    "height": 904
  },
  "articleUrl": null,
  "links": [
    {
      "label": "config.json",
      "href": "https://huggingface.co/tencent/Hy3-preview/blob/main/config.json",
      "title": "tencent/Hy3-preview"
    },
    {
      "label": "License",
      "href": "https://huggingface.co/tencent/Hy3-preview/blob/main/LICENSE",
      "title": null
    },
    {
      "label": "Tech report",
      "href": "https://www.tencent.com/en-us/articles/2202320.html",
      "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": "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": 17,
      "profile": {
        "Scientific": 3.6
      },
      "url": "https://artificialanalysis.ai/models/hy3-non-reasoning"
    }
  },
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-tencent-hy3-preview-295b-a21b",
  "family": "moe",
  "url": "https://llm-atlas-7or.pages.dev/models/tencent-hy3-preview-295b-a21b/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/tencent-hy3-preview-295b-a21b.md"
}