Autonomous Web Navigation on WebArena-Infinity
43.7Success RateQwen3.5 Plus
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3.5 PlusDeployment Mode=teacher-free test-time policy2026.06 | 43.7 | 30.49 | — | |
| SFT collector with SRCDeployment Mode=teacher-assisted data collection2026.06 | 42.5 | 29.58 | 12.01 | |
| Kimi K2.5Deployment Mode=teacher-free test-time policy2026.06 | 39.7 | 30.2 | — | |
| Expert SFT + SRCDeployment Mode=teacher-free test-time policy2026.06 | 35 | 28.1 | — | |
| LEAP-style post-hoc correctionDeployment Mode=teacher-assisted data collection2026.06 | 31.8 | 27.42 | 27.42 | |
| Expert SFTDeployment Mode=teacher-free test-time policy2026.06 | 25.3 | 12.06 | — | |
| OEC-style random switchDeployment Mode=teacher-assisted data collection2026.06 | 20.4 | 17.82 | 10.09 | |
| Base modelDeployment Mode=teacher-free test-time policy2026.06 | 15.8 | 29.83 | — |