agent evals / gilded-ocelot
Qwen/Qwen3.8-27B
| release | pass rate | trials | finished |
|---|---|---|---|
| acp-eval | 100.0% | 30 | 36d ago |
| acp-eval-quick | 100.0% | 3 | 36d ago |
per-task — acp-eval
| task | pass rate |
|---|---|
| 00-smoke | 100.0% |
| 01-long-output | 100.0% |
| 02-hanging-command | 100.0% |
| 03-binary-unicode | 100.0% |
| 04-multi-step-state | 100.0% |
| 05-recover-from-failure | 100.0% |
| 06-grep-edit-multifile | 100.0% |
| 07-install-dependency | 100.0% |
| 08-write-and-run-tests | 100.0% |
| 09-fix-cli-bug | 100.0% |