Alibaba (Qwen)
Qwen2.5-Coder-32B-Instruct benchmark results
6 results on 4 benchmarks from 4 sources.
Open weights Released: Nov 21, 2024 Updated Oct 22, 2026
Also listed as: qwen2.5-coder-32b-instruct · TTS(Bo8) · Skywork-SWE-32B + TTS(Bo8) · Qwen2.5 Coder 32B Instruct · Skywork-SWE-32B · Qwen2.5-Coder 32B Instruct · Qwen/Qwen2.5-Coder-32B-Instruct
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1271 | 309 | Overall, style control on; 5432 votes; 95% CI 1262–1279“qwen2.5-coder-32b-instruct” | Official leaderboard | Oct 8, 2026 |
| Open LLM Leaderboard v2 | 39.9% | 276 | Average of 6 normalised benchmarks; float16; 32.764B params“Qwen/Qwen2.5-Coder-32B-Instruct” | Official leaderboard | Mar 13, 2025 |
| HumanEval / HumanEval+ | 92.1% | 4 | HumanEval (base tests), pass@1 greedy | Official leaderboard | Dec 26, 2024 |
| HumanEval / HumanEval+ | 87.2% | 3 | HumanEval+ (extended tests), pass@1 greedy | Official leaderboard | Dec 26, 2024 |
| SWE-bench Verified | 47.0% | 70 | agent: Skywork-SWE-32B | Official leaderboard | Jun 16, 2025 |
| SWE-bench Verified | 9.0% | 105 | agent: mini-SWE-agent | Official leaderboard | Aug 3, 2025 |
FAQ
Where do these Qwen2.5-Coder-32B-Instruct scores come from?
From EvalPlus leaderboard, LMArena leaderboard dataset, Hugging Face Open LLM Leaderboard v2 (archived), and SWE-bench leaderboards (swebench.com). Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.