Qwen 1.5 7B: benchmark results
Alibaba's 7B model from the Qwen1.5 series (February 2024), adding 32K context and stronger multilingual performance over the original Qwen. Provider: Alibaba. Released 2024-02-04. Access: Open.
Aggregate Rating 1372 ± 10, rank #1327 of 1656 rated models, from 366 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM Lite - LegalBench - International Citizenship Questions | 63.3 | Exact Match (%) | 92.2 |
| HELM Lite - MATH Level 1 - Geometry | 68.42 | Equivalent (CoT) (%) | 82.2 |
| HELM NaturalQuestions (Open) | 74.92 | F1 (%) | 75.6 |
| Open Chinese LLM - WinoGrande | 65.19 | Accuracy (%) | 71.6 |
| Open Chinese LLM - C-Eval Semantic | 75.21 | Accuracy (%) | 65.4 |
| Open LLM Leaderboard v1 - GSM8K | 53.53 | Accuracy (%) (5-shot) | 65.2 |
| Open Arabic LLM - Aratrust Offensive | 84.06 | Accuracy (%) | 63.4 |
| Open PL LLM - PSC (multiple choice, 5-shot) | 69.48 | Binary F1 (%) | 60.2 |
| Open Arabic LLM - Arabic MMLU Accounting (University) | 55.41 | Accuracy (%) | 59 |
| Open Chinese LLM - CMMLU | 54.71 | Accuracy (%) | 58.3 |
| CMMLU | 72.5 | 5-shot Avg Accuracy (%) | 57.5 |
| Open Japanese LLM - Xlsum JA Rouge1 | 26.65 | Score (%) | 57.5 |
Interactive version: theaggregate.ai/model?slug=qwen-1-5-7b · How It Works · Data refreshed daily, snapshot 2026-10-20.