ministral-3B-2512: benchmark results
Mistral Ministral 3B 2512 checkpoint row. Provider: Mistral. Released 2025-12-02. Access: Open.
Aggregate Rating 1462 ± 13, rank #904 of 1635 rated models, from 218 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SpeechMap Compliance | 85.1 | % Requests Completed | 84.3 |
| SWE-Arena | 1002 | Elo Score | 81.9 |
| Evals for Every Language - Ewe (ee) | 29.92 | Average Score (%) | 53.2 |
| Evals for Every Language - Nuer (nus) | 26.22 | Average Score (%) | 50.8 |
| Evals for Every Language - Shan (shn) | 30.53 | Average Score (%) | 50.8 |
| Bongard Problems (Classic) | 8.8 | Accuracy (% correct) | 50 |
| Guesswork 2026-10 | 1.06 | MAE (z-score units) | 50 |
| Polish Knowledge Benchmark - Military | 75 | Accuracy (%) | 40.4 |
| Evals for Every Language - Sango (sg) | 31.86 | Average Score (%) | 38.1 |
| Konkur 1404 - Foreign Language | 34.62 | Accuracy (%, text-only) | 31.8 |
| Evals for Every Language - Oromo (om) | 31.24 | Average Score (%) | 31.7 |
| Polish Knowledge Benchmark - Cuisine | 71.71 | Accuracy (%) | 30.9 |
Interactive version: theaggregate.ai/model?slug=ministral-3b-2512 · How It Works · Data refreshed daily, snapshot 2026-11-05.