{
  "slug": "soofi-s-30b-a3b",
  "order": 85,
  "name": "Soofi S (30B-A3B)",
  "baseTitle": "Soofi S 30B-A3B",
  "company": "SOOFI Consortium",
  "country": "Germany",
  "flag": "🇩🇪",
  "date": "2026-07-10",
  "decoderType": "Hybrid MoE",
  "attention": "Mostly Mamba-2 with 6 GQA layers and no positional embeddings (NoPE)",
  "layerMix": "6 GQA + 23 Mamba-2 + 23 MoE",
  "scale": "31.6B total, 3.2B active (10.1% active)",
  "totalParams": 31600000000,
  "activeParams": 3200000000,
  "context": "1,048,576",
  "contextTokens": 1048576,
  "license": "Custom license (preview)",
  "kvCacheText": "6 KiB",
  "kvCacheKiB": 6,
  "kvCacheLevel": "Very low",
  "summary": "German-English base model that retrains the Nemotron 3 Nano architecture on 26.68T tokens with a fully documented data mixture.",
  "keyDetail": "Reuses the Nemotron 3 Nano architecture without modification, isolating the effect of a German-English data and training recipe.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/soofi.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/soofi.webp",
    "alt": "Figure 1: Soofi S 30B-A3B",
    "width": 944,
    "height": 960
  },
  "articleUrl": null,
  "links": [
    {
      "label": "config.json",
      "href": "https://huggingface.co/Soofi-Project/Soofi-S-Base/blob/main/config.json",
      "title": "Soofi-Project/Soofi-S-Base"
    },
    {
      "label": "License",
      "href": "https://huggingface.co/Soofi-Project/Soofi-S-Base#license-terms-of-use",
      "title": null
    },
    {
      "label": "Tech report",
      "href": "https://arxiv.org/abs/2607.09424",
      "title": null
    }
  ],
  "concepts": [
    {
      "slug": "nope",
      "label": "NoPE",
      "name": "No Positional Embeddings",
      "url": "https://llm-atlas-7or.pages.dev/concepts/nope/"
    },
    {
      "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": "hybrid-attention",
      "label": "Hybrid Attention",
      "name": "Hybrid Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/hybrid-attention/"
    }
  ],
  "aai": null,
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-soofi-s-30b-a3b",
  "family": "hybrid-moe",
  "url": "https://llm-atlas-7or.pages.dev/models/soofi-s-30b-a3b/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/soofi-s-30b-a3b.md"
}