Web Research on BrowseComp-ZH
29.1Pass@1Claude-4-Sonnet
Evaluation Results
| Method | Links | |
|---|---|---|
| Claude-4-Sonnet2025.12 | 29.1 | |
| THINKMERGEmodel=WebSailor-32B, N=42025.12 | 28.37 | |
| THINKMERGEmodel=WebSailor-32B, N=82025.12 | 27.34 | |
| THINKMERGEmodel=WebSailor-32B, N=22025.12 | 26.3 | |
| THINKMERGEmodel=WebSailor-7B, N=42025.12 | 24.91 | |
| THINKMERGEmodel=WebSailor-7B, N=82025.12 | 22.49 | |
| WebSailor-32Bvariant=Baseline2025.12 | 21.97 | |
| THINKMERGEmodel=WebSailor-7B, N=22025.12 | 21.45 | |
| ASearcher-Web-QwQparameters=32B2025.12 | 15.6 | |
| FoldAct-7Bp_drop=0.5, parameters=7B2025.12 | 15.2 | |
| FoldAct-7Bp_drop=0, parameters=7B2025.12 | 14.3 | |
| WebSailor-7Bparameters=7B2025.12 | 14.2 | |
| WebSailor-7Bvariant=Baseline2025.12 | 14.01 | |
| MiroThinker-8B-DPO-v0.1parameters=8B2025.12 | 13.6 | |
| THINKMERGEmodel=WebSailor-3B, N=22025.12 | 11.76 | |
| WebSailor-3Bvariant=Baseline2025.12 | 8.67 | |
| GPT-4.1-mini2025.12 | 8.3 | |
| THINKMERGEmodel=WebSailor-3B, N=42025.12 | 4.15 | |
| THINKMERGEmodel=WebSailor-3B, N=82025.12 | 2.77 |