Qwen2.5-Math-7B-Instruct: benchmark results
Alibaba's Qwen2.5-Math-7B-Instruct, the instruction-tuned version of Qwen2.5-Math-7B for solving English and Chinese math problems via chain-of-thought and tool-integrated reasoning. Provider: Alibaba. Released 2024-09-19. Access: Open.
Aggregate Rating 1319 ± 13, rank #1444 of 1656 rated models, from 125 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard - MATH Level 5 | 58.08 | Score | 99.6 |
| BankMathBench (English) - Basic | 61.2 | Exact-match accuracy (%) on the 559 English basic (single de | 86.4 |
| BankMathBench (Korean) - Basic | 46.2 | Exact-match accuracy (%) on the 559 Korean basic (single dep | 85 |
| BankMathBench (English) - Advanced | 2.5 | Exact-match accuracy (%) on the 399 English advanced (one pr | 81.8 |
| BankMathBench (Korean) - Advanced | 2 | Exact-match accuracy (%) on the 399 Korean advanced (one pro | 80 |
| Omni-MATH | 33.22 | Overall Accuracy (%) | 71.4 |
| BankMathBench (English) - Intermediate | 5 | Exact-match accuracy (%) on the 423 English intermediate (co | 65.9 |
| BankMathBench (Korean) - Intermediate | 0.9 | Exact-match accuracy (%) on the 426 Korean intermediate (com | 65 |
| MERA - SimpleAr | 98.9 | EM (%) | 63.6 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 28.05 | Exact match, flexible extract (%) | 62.4 |
| U-MATH - Sequences & Series | 51.95 | Accuracy (%) | 54.5 |
| MERA - MathLogicQA | 47.86 | Accuracy (%) | 53.8 |
Interactive version: theaggregate.ai/model?slug=qwen2-5-math-7b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.