Qwen 2.5 Coder 7B Instruct: benchmark results
Alibaba's Apache-2.0 7B code specialist launched with Qwen2.5-Coder in September 2024, trained on 5.5T code-heavy tokens with 128K context. Provider: Alibaba. Released 2024-09-19. Access: Open.
Aggregate Rating 1458 ± 9, rank #951 of 1656 rated models, from 180 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| RealBench (RAG) - Level 2 | 9.49 | Test pass rate (%), passed tests over all tests of each task | 100 |
| Open Korean LLM - Ko-IFEval | 51.75 | Strict accuracy, prompt/instruction mean (%) | 94.8 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 56.79 | Exact match, flexible extract (%) | 90.8 |
| Open Portuguese LLM - FaQuAD NLI | 80.48 | Macro F1 (%) | 90 |
| Open Korean LLM - Ko-GPQA-Diamond | 26.77 | Normalized accuracy (%) | 88.9 |
| GSM8K | 86.7 | Accuracy (%) | 88.7 |
| Open LLM Leaderboard - MATH Level 5 | 37.16 | Score | 87.5 |
| Open Korean LLM - Ko-EQ-Bench | 36.26 | EQ-Bench score | 86.3 |
| Open Ko-LLM Leaderboard | 42.19 | Average (%) | 85 |
| Open Korean LLM Leaderboard | 42.19 | Average Score (%) | 85 |
| MERA - RCB | 58.45 | Accuracy (%) | 84.3 |
| MERA - BPS | 99.4 | Accuracy (%) | 83.8 |
Interactive version: theaggregate.ai/model?slug=qwen-2-5-coder-7b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.