{
  "id": "mistral-7b-v0-1",
  "name": "Mistral 7B (v0.1)",
  "identifiers": {
    "huggingface": [
      "mistralai/Mistral-7B-v0.1"
    ]
  },
  "developer": "Mistral AI",
  "release_date": {
    "value": "2023-09-27",
    "source": "https://mistral.ai/news/announcing-mistral-7b",
    "status": "recorded"
  },
  "weights_status": "open",
  "availability": {
    "value": "available",
    "checked": "2026-09-24",
    "source": "https://huggingface.co/mistralai/Mistral-7B-v0.1"
  },
  "license": {
    "value": "Apache 2.0",
    "source": "https://mistral.ai/news/announcing-mistral-7b",
    "status": "recorded"
  },
  "architecture": {
    "family": "decoder_only",
    "n_layers": {
      "value": 32,
      "source": "https://huggingface.co/mistralai/Mistral-7B-v0.1/raw/27d67f1b5f57dc0953326b2601d68371d40ea8da/config.json",
      "status": "recorded"
    },
    "hidden_size": {
      "value": 4096,
      "source": "https://huggingface.co/mistralai/Mistral-7B-v0.1/raw/27d67f1b5f57dc0953326b2601d68371d40ea8da/config.json",
      "status": "recorded"
    },
    "n_heads": {
      "value": 32,
      "source": "https://huggingface.co/mistralai/Mistral-7B-v0.1/raw/27d67f1b5f57dc0953326b2601d68371d40ea8da/config.json",
      "status": "recorded"
    },
    "vocab_size": {
      "value": 32000,
      "source": "https://huggingface.co/mistralai/Mistral-7B-v0.1/raw/27d67f1b5f57dc0953326b2601d68371d40ea8da/config.json",
      "status": "recorded"
    },
    "positional_encoding": {
      "value": "rotary (RoPE)",
      "source": "https://github.com/mistralai/mistral-src/blob/c18d5b9166ffa16ce58827346ed56a28d2876e7e/one_file_ref.py",
      "status": "recorded",
      "note": "Not named in the paper; from Mistral's release-day reference code."
    },
    "context_length": {
      "value": 8192,
      "source": "https://arxiv.org/abs/2310.06825",
      "status": "recorded",
      "note": "Paper Table 1 context_len. Config max_position_embeddings is 32768, a loader limit, not the described design value."
    },
    "n_kv_heads": {
      "value": 8,
      "source": "https://huggingface.co/mistralai/Mistral-7B-v0.1/raw/27d67f1b5f57dc0953326b2601d68371d40ea8da/config.json",
      "status": "recorded"
    },
    "sliding_window": {
      "value": 4096,
      "source": "https://huggingface.co/mistralai/Mistral-7B-v0.1/raw/27d67f1b5f57dc0953326b2601d68371d40ea8da/config.json",
      "status": "recorded"
    },
    "note": "Paper: 'Compared to Llama, it introduces a few changes' (GQA, sliding window, rolling buffer cache). Effective attention span per layer is the 4096 sliding window.",
    "normalization": {
      "value": "RMSNorm, pre-norm",
      "source": "https://github.com/mistralai/mistral-src/blob/c18d5b9166ffa16ce58827346ed56a28d2876e7e/one_file_ref.py",
      "status": "recorded",
      "note": "Not named in the paper."
    },
    "activation": {
      "value": "SiLU in a gated feed-forward unit (SwiGLU)",
      "source": "https://github.com/mistralai/mistral-src/blob/c18d5b9166ffa16ce58827346ed56a28d2876e7e/one_file_ref.py",
      "status": "recorded",
      "note": "Not named in the paper; w2(silu(w1(x)) * w3(x))."
    },
    "ffn_hidden_size": {
      "value": 14336,
      "source": "https://arxiv.org/abs/2310.06825",
      "status": "recorded"
    }
  },
  "training_data": {
    "value": null,
    "source": "https://mistral.ai/news/announcing-mistral-7b",
    "status": "not_recorded",
    "note": "Neither the release post nor the paper discloses the pretraining corpus."
  },
  "techniques": [
    "transformer-decoder",
    "rotary-position-embedding",
    "rmsnorm",
    "swiglu",
    "grouped-query-attention",
    "sliding-window-attention"
  ],
  "primary_sources": [
    "https://arxiv.org/abs/2310.06825",
    "https://mistral.ai/news/announcing-mistral-7b",
    "https://huggingface.co/mistralai/Mistral-7B-v0.1/raw/27d67f1b5f57dc0953326b2601d68371d40ea8da/config.json"
  ],
  "record_history": [
    {
      "date": "2026-09-24",
      "change": "created from primary sources (Phase 1 seed, batch 1)",
      "by": "wilson-pruitt + claude"
    },
    {
      "date": "2026-09-24",
      "change": "context_length corrected 32768->8192 (paper); norm/activation/RoPE sourced to reference code; ffn width added (narrative/exercises/01-gpt2-xl-vs-mistral-7b.md)",
      "by": "wilson-pruitt + claude"
    },
    {
      "date": "2026-09-24",
      "change": "availability checked and recorded",
      "by": "wilson-pruitt + claude"
    }
  ]
}