Mistral AI releases Mistral Small 3
AIMade by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made
Released Mistral Small 3 — a latency-optimized 24B model, with both pretrained and instruction-tuned checkpoints under Apache 2.0
Why it matters
Mistral recommitted to Apache 2.0 for its general-purpose models while moving away from MRL-licensed ones; company-reported over 81% on MMLU at 150 tokens/s
Source — Primary sources first; secondary reports are labelled “Press”
Officialmistral.aihttps://mistral.ai/news/mistral-small-3Added in Week 41, 2026 · See that week
On the map — ⑨ Open-Weight Lineage line
- Previous stopLlama 3.2
- This recordMistral Small 3
- Next stopKanana Nano 2.1B
In the model family tree — Mistral family · 13 versions
- Earlier versionMistral Large 2
- This recordMistral Small 3
- Later versionMistral 3