Voxtral-Mini-3B-2507: benchmark results
Provider: Mistral. Released 2025-07-01. Access: Open.
Aggregate Rating 1446 ± 15, rank #1008 of 1635 rated models, from 201 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SEA-SpeechBench - Speech Translation (English Prompt) | 19.98 | BLEU (0-100): corpus BLEU of the English translation of Sout | 92.9 |
| EuroEval Finnish NLU - Tydiqa FI | 73.4 | Character F1 (%) | 91.2 |
| SEA-SpeechBench - Age Recognition (English Prompt) | 37.26 | Macro-F1 (%; speaker age group (teens, adults, seniors) from | 85.7 |
| EuroEval English NLU - SQuAD | 84.74 | Character F1 (%) | 85.6 |
| EuroEval German NLU - Germanquad | 63.91 | Character F1 (%) | 85.6 |
| EuroEval Lithuanian NLU - MultiWikiQA LT | 67.8 | Character F1 (%) | 85.6 |
| EuroEval Icelandic NLU - NQII | 54.94 | Character F1 (%) | 84.9 |
| EuroEval Spanish NLU - MLQA ES | 65.46 | Character F1 (%) | 84.6 |
| EuroEval Polish NLU - PoQuAD | 55.15 | Character F1 (%) | 81.5 |
| EuroEval French NLU - FQuAD | 70.85 | Character F1 (%) | 80.4 |
| EuroEval Latvian NLU - MultiWikiQA LV | 64.57 | Character F1 (%) | 78.9 |
| SEA-SpeechBench - Speech Translation (SEA Prompt) | 18.15 | BLEU (0-100): corpus BLEU of the English translation of Sout | 78.6 |
Interactive version: theaggregate.ai/model?slug=voxtral-mini-3b-2507 · How It Works · Data refreshed daily, snapshot 2026-10-15.