Qwen 2.5 7B: benchmark results
Alibaba's Apache-2.0 7B base model from the Qwen2.5 series (September 2024), pretrained on 18T tokens with 128K context. Provider: Alibaba. Released 2024-09-19. Access: Open.
Aggregate Rating 1462 ± 6, rank #922 of 1656 rated models, from 605 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Act2Intention - Intention Prediction Category Accuracy | 43.5 | Correct next-intention category (%) | 100 |
| LA Leaderboard - MGSM Catalan | 25.6 | Exact Match (%) | 100 |
| LA Leaderboard - MGSM Galician | 25.2 | Exact Match (%) | 100 |
| LA Leaderboard - XQuAD Catalan | 80.03 | F1 (%) | 100 |
| LA Leaderboard - XQuAD Spanish | 80.05 | F1 (%) | 100 |
| Relevance as a Vulnerability | 3.45 | Agent (self-reported) | 100 |
| LA Leaderboard - Belebele Galician | 86.44 | Accuracy (%) | 97.1 |
| LA Leaderboard - Belebele Spanish | 88.67 | Accuracy (%) | 97.1 |
| LA Leaderboard - CatalanQA | 83.39 | F1 (%) | 96.9 |
| LA Leaderboard - AQuAS | 70.73 | Semantic Answer Similarity (x100) | 95.6 |
| NormViz-Bench | 23 | Group Accuracy (%) | 95 |
| LA Leaderboard - Belebele Catalan | 85.56 | Accuracy (%) | 94.1 |
Interactive version: theaggregate.ai/model?slug=qwen-2-5-7b · How It Works · Data refreshed daily, snapshot 2026-10-20.