{
  "slug": "nemotron-3-nano-4b",
  "order": 56,
  "name": "Nemotron 3 Nano (4B)",
  "baseTitle": "Nemotron 3 Nano 4B",
  "company": "NVIDIA",
  "country": "United States",
  "flag": "🇺🇸",
  "date": "2026-03-16",
  "decoderType": "Dense hybrid",
  "attention": "GQA with only 4 attention layers",
  "layerMix": "4 GQA + 21 Mamba-2 + 17 FFN",
  "scale": "4B parameters",
  "totalParams": 4000000000,
  "activeParams": 4000000000,
  "context": "262,144",
  "contextTokens": 262144,
  "license": "NVIDIA Nemotron Open Model License",
  "kvCacheText": "16 KiB",
  "kvCacheKiB": 16,
  "kvCacheLevel": "Very low",
  "summary": "Compact on-device hybrid that compresses Nemotron Nano 9B v2 into a mostly Mamba-2 stack with only four attention layers.",
  "keyDetail": "Uses a 42-layer stack with 21 Mamba-2 blocks, 17 ReLU² FFNs, and just 4 GQA layers.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/nemotron-3-nano-4b.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/nemotron-3-nano-4b.webp",
    "alt": "Figure 1: Nemotron 3 Nano 4B",
    "width": 960,
    "height": 855
  },
  "articleUrl": null,
  "links": [
    {
      "label": "config.json",
      "href": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16/blob/main/config.json",
      "title": "nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16"
    },
    {
      "label": "License",
      "href": "https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-nemotron-open-model-license/",
      "title": null
    },
    {
      "label": "Tech report",
      "href": "https://huggingface.co/blog/nvidia/nemotron-3-nano-4b",
      "title": null
    }
  ],
  "concepts": [
    {
      "slug": "gqa",
      "label": "GQA",
      "name": "Grouped-Query Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/gqa/"
    },
    {
      "slug": "hybrid-attention",
      "label": "Hybrid Attention",
      "name": "Hybrid Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/hybrid-attention/"
    }
  ],
  "aai": {
    "v5": {
      "label": "AA Intelligence Index v5.0",
      "total": 1.8,
      "profile": {
        "Knowledge Work": 0.2,
        "Coding": 0,
        "Science": 0,
        "General": 7.1
      },
      "url": "https://artificialanalysis.ai/models/nvidia-nemotron-3-nano-4b"
    },
    "v4": {
      "label": "AA Intelligence Index v4.3.2",
      "total": 4.8,
      "profile": {
        "General": 8.8,
        "Scientific": 2.4,
        "Coding": 8,
        "Agents": 0.1
      },
      "url": "https://artificialanalysis.ai/models/nvidia-nemotron-3-nano-4b"
    }
  },
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-nemotron-3-nano-4b",
  "family": "hybrid",
  "url": "https://llm-atlas-7or.pages.dev/models/nemotron-3-nano-4b/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/nemotron-3-nano-4b.md"
}