Multi-agent Planning and Composition on TaskBench Dailylife APIs (test)
95.01PrecisionAutoMAS
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| AutoMASStrategy=ReAct Style, Critique=true, Iterative Correction=true2026.05 | 95.01 | 88.45 | 91.61 | 89.13 | 0.1031 | 90.34 | 62.12 | 79.54 |