{
  "slug": "zaya1-8b",
  "order": 63,
  "name": "ZAYA1-8B (8.4B)",
  "baseTitle": "ZAYA1-8B",
  "company": "Zyphra",
  "country": "United States",
  "flag": "🇺🇸",
  "date": "2026-05-06",
  "decoderType": "Sparse MoE",
  "attention": "CCA with 4:1 GQA, RoPE, and Q/K L2 norm",
  "layerMix": "40 CCA/GQA attention + 40 top-1 MoE FFN entries",
  "scale": "8.4B total, 760M active/token (9% active)",
  "totalParams": 8400000000,
  "activeParams": 760000000,
  "context": "131,072",
  "contextTokens": 131072,
  "license": "Apache License 2.0",
  "kvCacheText": "40 KiB",
  "kvCacheKiB": 40,
  "kvCacheLevel": "Low",
  "summary": "Zyphra's compact reasoning MoE pairs CCA/GQA attention updates with top-1 routed expert FFN updates to keep fewer than 1B parameters active per token.",
  "keyDetail": "Key detail is CCA/GQA attention, which performs attention in a compressed latent space while keeping a 4:1 GQA layout. Sidenote: the config lists 80 alternating layer entries, equivalent to 40 CCA/GQA attention updates paired with 40 top-1 MoE FFN updates.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/zaya1-8b.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/zaya1-8b.webp",
    "alt": "Figure 11: ZAYA1-8B",
    "width": 960,
    "height": 958
  },
  "articleUrl": "https://magazine.sebastianraschka.com/p/recent-developments-in-llm-architectures#%C2%A74-compressed-convolutional-attention-zaya1-8b",
  "links": [
    {
      "label": "View in article",
      "href": "https://magazine.sebastianraschka.com/p/recent-developments-in-llm-architectures#%C2%A74-compressed-convolutional-attention-zaya1-8b",
      "title": null
    },
    {
      "label": "config.json",
      "href": "https://huggingface.co/Zyphra/ZAYA1-8B/blob/main/config.json",
      "title": "Zyphra/ZAYA1-8B"
    },
    {
      "label": "Tech report",
      "href": "https://arxiv.org/abs/2605.05365",
      "title": null
    }
  ],
  "concepts": [
    {
      "slug": "compressed-convolutional-attention",
      "label": "CCA",
      "name": "Compressed Convolutional Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/compressed-convolutional-attention/"
    },
    {
      "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/"
    }
  ],
  "aai": null,
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-zaya1-8b",
  "family": "moe",
  "url": "https://llm-atlas-7or.pages.dev/models/zaya1-8b/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/zaya1-8b.md"
}