Mistral-7B-v0.3: benchmark results
Mistral AI's 7B base model, an update to v0.2 with an extended 32,768-token vocabulary, released under Apache 2.0. Provider: Mistral. Released 2024-05-22. Access: Open.
Aggregate Rating 1364 ± 12, rank #1355 of 1656 rated models, from 219 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LA Leaderboard - Summarization Galician | 7.55 | BLEU | 100 |
| LA Leaderboard - caBREU Catalan | 19.36 | BLEU | 98.5 |
| LA Leaderboard - CoQCat | 74.85 | F1 (%) | 95.4 |
| BOLD | 0.74 | BOLD score | 92.9 |
| LA Leaderboard - WNLI Basque | 59.15 | Accuracy (%) | 92.6 |
| LA Leaderboard - XL-Sum Spanish | 2.88 | BLEU | 90.3 |
| LA Leaderboard - TE-ca | 51.2 | Accuracy (%) | 89.7 |
| LA Leaderboard - XQuAD Catalan | 71.7 | F1 (%) | 89.2 |
| LA Leaderboard - CatalanQA | 79.56 | F1 (%) | 86.2 |
| LA Leaderboard - XQuAD Spanish | 72.56 | F1 (%) | 86.2 |
| LA Leaderboard - PAWS-X Spanish | 65.3 | Accuracy (%) | 85.3 |
| LA Leaderboard - ARC Catalan | 59.7 | Accuracy (%) | 83.8 |
Interactive version: theaggregate.ai/model?slug=mistral-7b-v0-3 · How It Works · Data refreshed daily, snapshot 2026-10-20.