Benchmark
SWE-bench Multilingual
Per-language resolution rates differ widely, so an average hides which languages the model actually handles.
- Released
- 2025-04-22
- Categories
- coding_agent
- Openness
- unknown
- Source
- model_reports
- Reported scores
- 7
Reported scores
model_reports
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| DeepSeek-V4-Flash | DeepSeek | 73.3 | 2026-06-25 | evidence |
| DeepSeek-V4-Pro | DeepSeek | 69.8 | 2026-04-22 | evidence |
| DeepSeek-V4-Pro | DeepSeek | 74.1 | 2026-04-22 | evidence |
| DeepSeek-V4-Pro | DeepSeek | 76.2 | 2026-04-22 | evidence |
| GLM-5 | Z.ai | 73.3 | 2026-02-11 | evidence |
| Hy4 preview | Tencent | 82.9 | 2026-08-28 | evidence |
| Qwen3.5-397B-A17B | Qwen | 69.3 | 2026-02-16 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.