Agentic on SWE-bench Verified
73.8ScoreGLM-5
Evaluation results
| Method | Links | |
|---|---|---|
| GLM-5Thinking Mode=non-thinking2026.06 | 73.8 | |
| Ling-2.6-1T2026.06 | 72.2 | |
| GPT-5.4Reasoning Mode=non-reasoning2026.06 | 69.2 | |
| Kimi-K2.5Mode=Instant2026.06 | 66.8 | |
| DeepSeek-V3.2Thinking Mode=nothink2026.06 | 66.4 | |
| Ling-2.6-flash2026.06 | 61.2 | |
| Nemotron-3-Super 120B-A12BEvaluation Setting=non-reasoning2026.06 | 61 | |
| Audex 30B-A3BContext Length=1M2026.07 | 48.2 | |
| GPT-5.4-miniEvaluation Setting=non-reasoning2026.06 | 43.8 | |
| Qwen3-Omni 30B-A3B ThinkingContext Length=64K2026.07 | 23.4 |