Qwen 2.5 3B: benchmark results
Alibaba's 3B model from the Qwen2.5 series (September 2024), pretrained on up to 18T tokens and released under the non-commercial Qwen Research License. Provider: Alibaba. Released 2024-09-19. Access: Open.
Aggregate Rating 1403 ± 10, rank #1211 of 1656 rated models, from 421 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LA Leaderboard - MGSM Spanish | 18 | Exact Match (%) | 100 |
| LA Leaderboard - MGSM Galician | 16.4 | Exact Match (%) | 96.9 |
| LA Leaderboard - XQuAD Spanish | 77 | F1 (%) | 95.4 |
| LA Leaderboard - MGSM Catalan | 12.8 | Exact Match (%) | 93.8 |
| LA Leaderboard - RagQuAS | 80.86 | Semantic Answer Similarity (x100) | 92.6 |
| LA Leaderboard - XQuAD Catalan | 71.76 | F1 (%) | 90.8 |
| NormViz-Bench - Macro F1 | 44.6 | Macro F1 (%) | 90 |
| LA Leaderboard - PAWS Galician | 65.7 | Accuracy (%) | 89 |
| LA Leaderboard - WNLI Catalan | 63.38 | Accuracy (%) | 86 |
| Relevance as a Vulnerability | 3.22 | Agent (self-reported) | 85.7 |
| LA Leaderboard - Belebele Catalan | 80.22 | Accuracy (%) | 85.3 |
| LA Leaderboard - Belebele Galician | 79.78 | Accuracy (%) | 85.3 |
Interactive version: theaggregate.ai/model?slug=qwen-2-5-3b · How It Works · Data refreshed daily, snapshot 2026-10-20.