| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Nexus (Held-Out) | SHAD+RFT | Accuracy32 | 14 | 4mo ago | |
| T-eval (Held-Out) | SHAD+RFT | Accuracy71.8 | 14 | 4mo ago | |
| StableToolBench Held-In | SHAD+RFT | Pass Rate50.4 | 14 | 4mo ago | |
| ClawEval 114 T-family tasks | SFT ISETrace | p@137.7 | 4 | 1mo ago | |
| WorkBench | MemToolAgent | Analytics Success Rate92.59 | 4 | 1mo ago | |
| OmniGAIA | Score68.9 | 3 | 3mo ago |