Agent on SWE-bench Verified
72.8AccuracyGLM-5
Evaluation Results
| Method | Links | |
|---|---|---|
| GLM-5Scale Band=Frontier, Agent Harness=Mini-SWE-Agent2026.06 | 72.8 | |
| DeepSeek V3.2Evaluation Mode=Chat2025.12 | 72.1 | |
| GLM 4.6Evaluation Mode=Chat2025.12 | 68 | |
| LongCat-Flash Exp-ChatEvaluation Mode=Chat2025.12 | 63.2 | |
| LongCat-Flash ChatEvaluation Mode=Chat2025.12 | 60.4 | |
| OpenThinkerAgent-32BScale Band=32B, Agent Harness=terminus-22026.06 | 54 | |
| SWE-Lego-Qwen3-8BScale Band=8B, Agent Harness=OpenHands2026.06 | 37.6 | |
| SERA-8BScale Band=8B, Agent Harness=SERA-SWE-Agent2026.06 | 31.5 | |
| Nemotron-Terminal-8BScale Band=8B, Agent Harness=terminus-22026.06 | 22.1 |