| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| BFCL multi-turn V3 | Average Success Rate68 | 48 | 1mo ago | ||
| AppWorld | SeeUPO | Avg@463.6 | 25 | 4mo ago | |
| FTRL, BFCL, and ToolHop Combined | PACT | Average Success Rate36.9 | 18 | 1mo ago | |
| BFCL | PACT | Multi-Turn Success Rate27.5 | 18 | 1mo ago | |
| FTRL | PACT | Solve Rate (R)42.41 | 18 | 1mo ago | |
| Xbench | WebAnchor-30B | Pass@175.1 | 18 | 4mo ago | |
| BrowseComp-ZH | Pass@158.1 | 18 | 4mo ago | ||
| BrowseComp | Pass@150.9 | 18 | 4mo ago | ||
| GAIA | WebAnchor-30B | Pass@176.4 | 18 | 4mo ago | |
| τ³-airline 20-task (held-out) | SGCD | Pass@160.2 | 4 | 1mo ago | |
| AppWorld (test challenge) | SGCD | Task Goal Completion (TGC)27 | 3 | 1mo ago |