Software issue-style repair on QGP-SWE lite
91.7Success RateStandard
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| StandardAgent=Aider, Conditions=242026.05 | 91.7 | 2.67 | 8.3 | 0.173 | 0.653 | |
| UNITQGPAgent=Aider, Conditions=242026.05 | 91.7 | 2.67 | 8.3 | 0.176 | 0.667 | |
| UNITQGPAgent=LangGraph, Conditions=242026.05 | 41.7 | 1.46 | 58.3 | 0.027 | 0.294 | |
| UNITQGPAgent=Native, Conditions=242026.05 | 37.5 | 1.25 | 62.5 | 0.022 | 0.25 | |
| StandardAgent=Native, Conditions=242026.05 | 0 | 0 | 100 | 0 | 0 | |
| Verifier-gatedAgent=Native, Conditions=242026.05 | 0 | 0 | 100 | 0 | 0 | |
| StandardAgent=LangGraph, Conditions=242026.05 | 0 | 0 | 100 | 0 | 0 | |
| Verifier-gatedAgent=LangGraph, Conditions=242026.05 | 0 | 0 | 100 | 0 | 0 |