Web Navigation on Mind2Web Cross-website 1.0
47.1Success RatePolySkill (+Update)
Evaluation Results
| Method | Links | |
|---|---|---|
| PolySkill (+Update)Backbone=GLM-4.5, Training tasks=1009 tasks2025.10 | 47.1 | |
| ASI (+Update)Backbone=GLM-4.5, Training tasks=1009 tasks2025.10 | 46.2 | |
| PolySkillBackbone=GLM-4.5, Training tasks=1009 tasks2025.10 | 44.7 | |
| ASIBackbone=GLM-4.5, Training tasks=1009 tasks2025.10 | 44.5 | |
| BaselineBackbone=GLM-4.5, Training tasks=1009 tasks2025.10 | 44.2 | |
| PolySkill (+Update)Backbone=Qwen3-Coder-480B-A35B, Training tasks=1009 tasks2025.10 | 41.8 | |
| ASI (+Update)Backbone=Qwen3-Coder-480B-A35B, Training tasks=1009 tasks2025.10 | 40.1 | |
| BaselineBackbone=Qwen3-Coder-480B-A35B, Training tasks=1009 tasks2025.10 | 38.1 | |
| PolySkillBackbone=Qwen3-Coder-480B-A35B, Training tasks=1009 tasks2025.10 | 37.2 | |
| ASIBackbone=Qwen3-Coder-480B-A35B, Training tasks=1009 tasks2025.10 | 36.4 |