SWE-agent
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Failure prediction | SWE-agent (Transfer held-out) | ROC-AUC0.69 | 7 | |
| Failure prediction | SWE-agent (Within corpus) | ROC-AUC0.713 | 4 | |
| Software engineering task resolution | SWE-agent task distribution | Tokens per Success74,000 | 3 |
Showing 3 of 3 rows