Qwen 2.5 1.5B Instruct: benchmark results
Alibaba Qwen 2.5 1.5B instruction-tuned checkpoint. Provider: Alibaba. Released 2024-09-19. Access: Open.
Aggregate Rating 1340 ± 7, rank #1407 of 1656 rated models, from 829 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SLM Trustworthiness - Fairness | 71.82 | TrustLLM fairness accuracy (%; sub-task mean) | 100 |
| LA Leaderboard - ClinTreatES | 65.27 | Semantic Answer Similarity (x100) | 95.5 |
| RobustBench-TC - Authentication Error | 32.2 | Tool-call accuracy (%) under the AuthErr perturbation (trans | 94.1 |
| SLM Trustworthiness - Robustness | 69.38 | TrustLLM robustness accuracy (%; sub-task mean) | 90 |
| SLM Trustworthiness - TrustLLM Overall | 64.9 | TrustLLM score (%; mean of ethics, privacy, robustness and f | 90 |
| Structured Local-Deployment MCQ | 67.1 | Strict accuracy (%): share of the 1,085 items answered with | 87.5 |
| MERA Code - CodeCorrectness | 83.69 | Accuracy (%) | 86.2 |
| RobustBench-TC - Rate Limit | 31.7 | Tool-call accuracy (%) under the RateLim perturbation (trans | 82.4 |
| LA Leaderboard - PAWS-X Spanish | 64.05 | Accuracy (%) | 77.2 |
| LA Leaderboard - NoticIA | 22.07 | ROUGE-1 (%) | 75.4 |
| EuroEval English NLU - SQuAD | 83.18 | Character F1 (%) | 74.8 |
| Open LLM Leaderboard - MATH Level 5 | 22.05 | Score | 74.8 |
Interactive version: theaggregate.ai/model?slug=qwen-2-5-1-5b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.