{
  "slug": "ling-3-1-flash",
  "order": 109,
  "name": "Ling 3.1 Flash (560B)",
  "baseTitle": "Ling 3.1 Flash",
  "company": "InclusionAI",
  "country": "China",
  "flag": "🇨🇳",
  "date": "2026-10-13",
  "decoderType": "Sparse hybrid",
  "attention": "7:1 Kimi Delta Attention and gated MLA",
  "layerMix": "63 Kimi Delta Attention + 9 gated MLA + 1 MTP layer",
  "scale": "560B total, 25B active (4.5% active)",
  "totalParams": 560000000000,
  "activeParams": 25000000000,
  "context": "1M reported; 262,144 in config",
  "contextTokens": 1048576,
  "license": "MIT License",
  "kvCacheText": "10.1 KiB",
  "kvCacheKiB": 10.1,
  "kvCacheLevel": "Very low",
  "summary": "Hybrid MoE that extends Ling 3.0 Flash to 72 layers, with seven Kimi Delta Attention layers for every gated MLA layer.",
  "keyDetail": "Uses four initial dense FFN layers, top-8 routing over 512 experts plus one shared expert, and MLA with a 1,536-dimensional query bottleneck and 512-dimensional KV latent.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/ling-3-1-flash.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/ling-3-1-flash.webp",
    "alt": "Figure 1: Ling 3.1 Flash",
    "width": 960,
    "height": 946
  },
  "articleUrl": null,
  "links": [
    {
      "label": "config.json",
      "href": "https://huggingface.co/inclusionAI/Ling-3.1-flash/blob/main/config.json",
      "title": "inclusionAI/Ling-3.1-flash"
    },
    {
      "label": "Tech report",
      "href": "https://huggingface.co/inclusionAI/Ling-3.1-flash",
      "title": null
    }
  ],
  "concepts": [
    {
      "slug": "mla",
      "label": "MLA",
      "name": "Multi-Head Latent Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/mla/"
    },
    {
      "slug": "gated-attention",
      "label": "Gated Attention",
      "name": "Gated Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/gated-attention/"
    },
    {
      "slug": "hybrid-attention",
      "label": "Hybrid Attention",
      "name": "Hybrid Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/hybrid-attention/"
    },
    {
      "slug": "shortconv",
      "label": "ShortConv",
      "name": "Short Convolution",
      "url": "https://llm-atlas-7or.pages.dev/concepts/shortconv/"
    },
    {
      "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": null,
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-ling-3-1-flash",
  "family": "hybrid-moe",
  "url": "https://llm-atlas-7or.pages.dev/models/ling-3-1-flash/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/ling-3-1-flash.md"
}