Codesota · Models · Qwen2.5-Coder 32BAlibaba9 results · 8 benchmarks
Model card

Qwen2.5-Coder 32B.

Alibabaopen-source32B paramsDense Transformer

32B parameters. SOTA open-source code model at release (Nov 2024). Matches GPT-4o on HumanEval.

§ 01 · Benchmarks

Every benchmark Qwen2.5-Coder 32B has a recorded score for.

#BenchmarkArea · TaskMetricValueRankDateSource
01Bugs2FixComputer Code · Bug Detectionaccuracy76.8%#2/62024-09-19source ↗
02CodeSearchNetComputer Vision · Optical Character Recognitionbleu-423.4%#2/72024-09-19source ↗
03CrossCodeEvalComputer Code · Code Completionexact-match43.7%#2/62024-09-19source ↗
04TransCoder (GeeksForGeeks)Computer Code · Code Translationcomputational-accuracy86.3%#3/72024-09-19source ↗
05MBPPComputer Code · Code Generationpass@190.2%#5/192024-09-19source ↗
06HumanEvalComputer Code · Code Generationpass@192.7%#9/422025-03-01source ↗
07HumanEvalComputer Code · Code Generationpass@192.7%#9/422024-09-19source ↗
08SWE-BenchComputer Code · Code Generationresolve-rate55.4%#22/322025-06-01source ↗
09LiveCodeBenchComputer Code · Code Generationpass@147.8%#22/302024-03-12source ↗
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area

Where Qwen2.5-Coder 32B actually performs.

Computer Vision
1
benchmark
avg rank #2.0
Computer Code
7
benchmarks
avg rank #9.3
§ 03 · Papers

3 papers with results for Qwen2.5-Coder 32B.

  1. 2024-09-19· Computer Code· 6 results

    Qwen2.5-Coder Technical Report

  2. 2024-03-12· Computer Code· 1 result

    LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code

  3. 2023-10-10· Computer Code· 1 result

    SWE-bench: Can Language Models Resolve Real-World GitHub Issues?

    Carlos E. Jimenez, John Yang, Alexander Wettig, Shunyu Yao et al.
§ 04 · Related models

Other Alibaba models scored on Codesota.

Qwen2-VL 72B
4 results
Qwen2.5-72B-Instruct
72B params · 4 results
GOT-OCR2.0
3 results
Qwen 3 72B
72B params · 2 results
Qwen2.5-VL 32B
2 results
Qwen2.5-VL 72B
72B params · 2 results
Qwen 3 14B
14B params · 1 result
Qwen2-VL 7B
7B params · 1 result
§ 05 · Sources & freshness

Where these numbers come from.

arxiv
6
results
shadow-page-humaneval
1
result
swebench-leaderboard
1
result
official-leaderboard
1
result
9 of 9 rows marked verified. · first result 2024-03-12, latest 2025-06-01.