calme-3.2-baguette-3B: benchmark results
MaziyarPanahi's French-English fine-tune of Qwen2.5-3B, trained on his LegalKit and French instruction datasets. Provider: Other. Released 2024-11-07. Access: Open.
Unified ELO 1517 ± 1, rank #620 of 1392 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard - MATH Level 5 | 28.25 | Score | 80.5 |
| Open LLM Leaderboard - IFEval | 63.38 | Score | 75.5 |
| French LLM Leaderboard - IFEval FR | 46.35 | Score (%) | 63 |
| French LLM Leaderboard - GPQA FR | 30.22 | Score (%) | 60.2 |
| French LLM Leaderboard - Average | 34 | Average Score (%) | 57.4 |
| Open LLM Leaderboard - GPQA | 5.93 | Score | 49.7 |
| Open LLM Leaderboard - MMLU-Pro | 25.98 | Score | 47 |
| French LLM Leaderboard - BAC FR | 25.42 | Score (%) | 42.6 |
| Open LLM Leaderboard - MuSR | 8.6 | Score | 41.4 |
| Open LLM Leaderboard - BBH | 25.87 | Score | 40 |
Interactive version: theaggregate.ai/model?slug=calme-3-2-baguette-3b · How It Works · Data refreshed daily, snapshot 2026-09-05.