{
  "id": "mistral-small-4",
  "name": "Mistral Small 4",
  "developer": "Mistral AI",
  "exact_model_id": "mistralai/Mistral-Small-4-119B-2603",
  "verified_at": "2026-09-27",
  "evidence_level": "source-verified",
  "access": {
    "label": "Direct download",
    "gated": false
  },
  "license": {
    "name": "Apache 2.0",
    "commercial_use": "Apache 2.0 supports commercial and non-commercial use.",
    "eu_note": "No explicit EU territorial restriction identified in the reviewed model materials."
  },
  "architecture": {
    "parameters": "119B total / 6.5B active per token",
    "type": "MoE — 128 experts, 4 active",
    "context": "256K",
    "modalities": "Text + image → text"
  },
  "runtimes": [
    "Transformers",
    "vLLM"
  ],
  "hardware_note": "Official NVFP4 checkpoint can materially reduce memory versus BF16.",
  "primary_source": "https://huggingface.co/mistralai/Mistral-Small-4-119B-2603",
  "third_party_runtime_evidence": [],
  "reference_page": "https://openweightmodels.eu/model-mistral-small-4.html",
  "owm_view": "Mistral Small 4 is an unusually dense package of deployment features: multimodal input, reasoning and non-reasoning modes, function calling, a 256K context window, MoE efficiency and Apache 2.0 licensing. OWM sees the official NVFP4 checkpoint as especially important because it turns quantization from a community afterthought into a publisher-supported deployment path.",
  "seo_reference_verified_at": "2026-09-27",
  "change_history": "https://openweightmodels.eu/change-mistral-small-4.json"
}