Mistral 7B
Summary
Mistral 7B v0.1 is a 7-billion-parameter language model designed for performance and efficiency. The authors report that it outperforms Llama 2 13B across all evaluated benchmarks, and Llama 1 34B in reasoning, mathematics, and code generation. The model uses grouped-query attention (GQA) for faster inference and sliding window attention (SWA) to handle sequences of arbitrary length at reduced inference cost. They also provide an instruction-finetuned model, Mistral 7B Instruct, that they say surpasses Llama 2 13B Chat on human and automated benchmarks. The models are released under the Apache 2.0 license. [S1]
Claims
- id: C1 claim: 'Mistral 7B v0.1 is a 7-billion-parameter language model.' source: S1 evidence: "Mistral 7B v0.1, a 7-billion-parameter language model"
- id: C2 claim: 'The authors report it outperforms Llama 2 13B on all evaluated benchmarks.' source: S1 evidence: "Mistral 7B outperforms Llama 2 13B across all evaluated benchmarks"
- id: C3 claim: 'The models are released under the Apache 2.0 license.' source: S1 evidence: "Our models are released under the Apache 2.0 license"
Verification
- Code available: Comments point to https://mistral.ai/news/announcing-mistral-7b/, which returned HTTP 200 on 2026-11-12. The HTML also names https://github.com/mistralai/mistral-src, which returned HTTP 200 and redirected to https://github.com/mistralai/mistral-inference. Not run.
- Reproduced: not reproduced (2026-11-12). Code not run.
- Related work: Live notes 2307-09288-llama-2 and 2305-13245-gqa. This is the paper-verification note for Mistral 7B itself.
- Limits: Comparisons with Llama 2 13B and Llama 1 34B are author-reported on the paper's evaluation suite (v1, 10 Oct 2023). This note does not restate every benchmark score.
- License: CC BY 4.0 on the abs page (2026-11-12; icon by-4.0.png; http://creativecommons.org/licenses/by/4.0/). Submitted 10 Oct 2023 (v1). Sources: [S1] https://arxiv.org/abs/2310.06825, v1 (10 Oct 2023). checked_on: 2026-11-12.
When to apply
Use when comparing small open language models, or when tracing grouped-query and sliding-window attention in a 7B baseline.
Facts
- id: F1 claim: 'Mistral 7B v0.1 is a 7-billion-parameter language model that uses grouped-query attention and sliding window attention.' confidence: medium date: 2026-11-12 sources: [S1] license: CC BY 4.0
- id: F2 claim: 'The authors report that it outperforms Llama 2 13B across all evaluated benchmarks.' confidence: medium date: 2026-11-12 sources: [S1] license: CC BY 4.0
- id: F3 claim: 'The models are released under the Apache 2.0 license.' confidence: medium date: 2026-11-12 sources: [S1] license: CC BY 4.0
Sources
- [S1] arXiv 2310.06825, https://arxiv.org/abs/2310.06825. License: CC BY 4.0. v1 (10 Oct 2023). Accessed 2026-11-12.
License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Summary by iinventedlight (AI-assisted). Paper is CC BY 4.0.
License: CC-BY-4.0 · Summary by iinventedlight (AI-assisted). Source: Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch, and 15 coauthors (listed on the arXiv abs page), Mistral 7B (CC BY 4.0).
Cite this
'iinventedlight (AI-assisted)'. 2026. "Mistral 7B." iinventedlight Library, iil:note:8v23vdf1, version of 2026-11-12. https://iinventedlight.com/library/notes/2310-06825-mistral-7b/. License: CC-BY-4.0.
@misc{iil_note_8v23vdf1,
author = {{'iinventedlight (AI-assisted)'}},
title = {{Mistral 7B}},
howpublished = {iinventedlight Library},
year = {2026},
date = {2026-11-12},
url = {https://iinventedlight.com/library/notes/2310-06825-mistral-7b/},
note = {Permanent ID iil:note:8v23vdf1. License: CC-BY-4.0.},
}
Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.
For language models and agents: start at /llms.txt | /agents.md | /.well-known/agent-library.json. This page in Markdown: /library/notes/2310-06825-mistral-7b.md.