Named plus request share: mistralai is 2.6% of text requests (41.2M). Large 3 2512 is the capable flagship (AA 16). Medium 3.5 is newer and higher on AA (30) and has its own card.
This lab was not on the This Week LLM Leaderboard or the named Top Weekly rows we cite. No week-token figure is invented. Fetched 2026-08-19. Text request share for mistralai: 2.6% (41.2M requests) in the same window. That is request share, not tokens. OpenRouter This Week tokens are prompt plus completion. The percent next to a row is week-over-week change, not share of the week. Market share on this page is text request share by author, a different number.
Capable flagship. Official HF mistralai/Mistral-Large-3-675B-Instruct-2512. AA 16.
Newer sibling, higher on AA (30). Source: Artificial Analysis, 19 Aug 2026.
Mistral Small 4: 2026-08-19 OpenRouter row.
Ministral 3 14B 2512: 2026-08-19 OpenRouter row.
Ministral 3 8B 2512: 2026-08-19 OpenRouter row.
Ministral 3 3B 2512: 2026-08-19 OpenRouter row.
Mistral Medium 3.1: 2026-08-19 OpenRouter row.
Codestral 2508: 2026-08-19 OpenRouter row.
Mistral Small 3.2 24B: 2026-08-19 OpenRouter row.
Mistral Medium 3: 2026-08-19 OpenRouter row.
Mistral Small 3.1 24B: 2026-08-19 OpenRouter row.
Saba: 2026-08-19 OpenRouter row.
Mistral Small 3: 2026-08-19 OpenRouter row.
Mistral Large 2407: 2026-08-19 OpenRouter row.
Capable Large. Open weights. AA 16.
Current Medium. AA 30. Closed on the router.
Mistral Small 4: 2026-08-19 OpenRouter row.
Ministral 3 14B 2512: 2026-08-19 OpenRouter row.
Ministral 3 8B 2512: 2026-08-19 OpenRouter row.
Ministral 3 3B 2512: 2026-08-19 OpenRouter row.
Mistral Medium 3.1: 2026-08-19 OpenRouter row.
Codestral 2508: 2026-08-19 OpenRouter row.
Mistral Small 3.2 24B: 2026-08-19 OpenRouter row.
Mistral Medium 3: 2026-08-19 OpenRouter row.
Mistral Small 3.1 24B: 2026-08-19 OpenRouter row.
Saba: 2026-08-19 OpenRouter row.
Mistral Small 3: 2026-08-19 OpenRouter row.
Mistral Large 2407: 2026-08-19 OpenRouter row.
This hub lists every published card. Start with Mistral Large 3.
Ranked view: Best coding models: the independent leaderboard puts every Mistral card with an independent coding row next to every other lab on one board, with the confidence intervals left visible.
The app is free with your own keys. Plus adds Continuum-hosted inference and the ids we actually list on each card.