Magistral Small 1.2: benchmark results
Mistral's September 2025 Magistral Small 1.2, a 24B multimodal reasoning checkpoint. Provider: Mistral. Released 2025-09-17. Access: Open.
Aggregate Rating 1564 ± 16, rank #562 of 1635 rated models, from 287 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval Swedish NLU - Multi Wiki QA SV | 82.4 | Character F1 (%) | 99.6 |
| EuroEval Slovak NLU - Multi Wiki QA SK | 74.29 | Character F1 (%) | 99 |
| EuroEval Danish NLU - Multi Wiki QA DA | 82.72 | Character F1 (%) | 98.9 |
| EuroEval Faroese NLU - FoQA | 79.56 | Character F1 (%) | 98.7 |
| EuroEval Italian NLU - SQuAD IT | 76.71 | Character F1 (%) | 98.6 |
| EuroEval Norwegian NLU - Norquad | 77.06 | Character F1 (%) | 98.4 |
| EuroEval Latvian NLU - MultiWikiQA LV | 71.87 | Character F1 (%) | 97.8 |
| EuroEval English | 77.21 | Average Score (%) | 97.6 |
| EuroEval Lithuanian NLU - MultiWikiQA LT | 73.58 | Character F1 (%) | 97.4 |
| EuroEval Spanish NLU - MLQA ES | 69.43 | Character F1 (%) | 97.3 |
| EuroEval English NLU - SQuAD | 88.59 | Character F1 (%) | 97.2 |
| EuroEval Slovak NLU | 61.56 | NLU Average Score (%) | 97 |
Interactive version: theaggregate.ai/model?slug=magistral-small-1-2 · How It Works · Data refreshed daily, snapshot 2026-10-15.