Qwen 1.5 72B: benchmark results
72B base model of Alibaba's open-weights Qwen1.5 series (February 2024), a multilingual release with 32K context. Provider: Alibaba. Released 2024-02-04. Access: Open.
Aggregate Rating 1511 ± 14, rank #753 of 1656 rated models, from 149 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - MMLU | 77.2 | Accuracy (%) (5-shot) | 98.4 |
| Open PL LLM - PSC (multiple choice, 5-shot) | 94.27 | Binary F1 (%) | 98.4 |
| HELM MMLU - High School Government And Politics | 98.96 | EM | 98.1 |
| Open PL LLM - KLEJ NER (multiple choice, 5-shot) | 59.77 | Accuracy (%) | 95.6 |
| Open PL LLM - PSC (generative, 5-shot) | 98.35 | Binary F1 (%) | 94.6 |
| Open Chinese LLM - HellaSwag | 70.33 | Accuracy (%) | 94.1 |
| Open Chinese LLM - C-Eval Semantic | 90.13 | Accuracy (%) | 93.5 |
| Open Chinese LLM - GSM8K | 65.28 | Accuracy (%) | 89.9 |
| Open Chinese LLM Leaderboard | 68.14 | Official Nonzero-Task Mean (%) | 89.9 |
| ShoppingMMLU | 67.39 | Average Score (%) | 88.5 |
| Open Chinese LLM - WinoGrande | 68.82 | Accuracy (%) | 87.9 |
| Open PL LLM - PolEmo2-OUT (generative, 5-shot) | 78.34 | Accuracy (%) | 87.9 |
Interactive version: theaggregate.ai/model?slug=qwen-1-5-72b · How It Works · Data refreshed daily, snapshot 2026-10-20.