| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Misalignment Detection | WorkBench (n=326) | Average IR52 | 30 | |
| Misalignment Detection | Workbench n=62 (general) | Error Rate (ER)0.3 | 30 | |
| Multi-agent system task solving | workbench | Accuracy82.4 | 21 | |
| Job and Skill Intelligence Tasks | WorkBench (test) | Job2Skill MAP17.9 | 9 | |
| Ranking | WorkBench | Latency (ms)13.2 | 9 | |
| Agent Tool-Use | WorkBench | Analytics Success Rate92.59 | 4 | |
| General Purpose Tool Use | WorkBench | Analytics Success Rate86.11 | 4 | |
| Multi-tool Coordination | Workbench | Beta_Ec x T0.267 | 1 |