Web Navigation on OSWorld (All)
27Success Rate (All)Base Agent + Verifier, SGV
Evaluation results
| Method | Links | |
|---|---|---|
| Base Agent + Verifier, SGVVerification=SGV2025.07 | 27 | |
| Base Agent + Verifier, no SGVVerification=no SGV2025.07 | 24 | |
| Base Agent2025.07 | 22 |
| Method | Links | |
|---|---|---|
| Base Agent + Verifier, SGVVerification=SGV2025.07 | 27 | |
| Base Agent + Verifier, no SGVVerification=no SGV2025.07 | 24 | |
| Base Agent2025.07 | 22 |