Qwen 2 0.5B: benchmark results
Alibaba's Qwen2-0.5B base language model, part of the Qwen2 series spanning 0.5 to 72 billion parameters plus a Mixture-of-Experts variant, using grouped-query attention. Provider: Alibaba. Released 2024-06-06. Access: Open.
Aggregate Rating 1242 ± 17, rank #1547 of 1656 rated models, from 131 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 35.71 | Accuracy (%) (5-shot) | 51.2 |
| Open Japanese LLM - Mbpp Pylint Check | 30.72 | Score (%) | 50.4 |
| MEDIC Benchmark | 62.2 | MEDIC Public Table Average | 48.4 |
| Open Japanese LLM - CG | 6.02 | Score (%) | 46.5 |
| MERA - ruHumanEval | 3.6 | pass@1 (%) | 32.3 |
| Open PL LLM - PSC (generative, 5-shot) | 80.02 | Binary F1 (%) | 29.7 |
| MERA - ruCodeEval | 0.55 | pass@1 (%) | 26.7 |
| Open LLM Leaderboard - MuSR | 37.52 | Score | 26.7 |
| MERA - ruDetox | 9.13 | Joint Score (%) | 24.9 |
| Open PL LLM - CBD (generative, 5-shot) | 16.57 | Macro F1 (%) | 24.5 |
| Open PL LLM - CBD (multiple choice, 5-shot) | 15.89 | Macro F1 (%) | 24.1 |
| Open LLM Leaderboard v1 - MMLU | 45.17 | Accuracy (%) (5-shot) | 23.8 |
Interactive version: theaggregate.ai/model?slug=qwen-2-0-5b · How It Works · Data refreshed daily, snapshot 2026-10-20.