Computer-use on OSWorld 100-step (Headline SR)
72.6Headline SRAgent S3 (bBoN)
Evaluation Results
| Method | Links | |
|---|---|---|
| Agent S3 (bBoN)Grounding=Behavior-narrative judge, Cost axis=N=10 rollouts (∼10×)2026.05 | 72.6 |
Loading the SOTA2 catalog…
| Method | Links | |
|---|---|---|
| Agent S3 (bBoN)Grounding=Behavior-narrative judge, Cost axis=N=10 rollouts (∼10×)2026.05 | 72.6 |