| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| ALFWorld | RSPO+GiGPO | Pick Success Rate98.9 | 36 | 18d ago | |
| TRIP-Bench (Hard LIT) | Loose Success4,400 | 32 | 4mo ago | ||
| TRIP-Bench Overall | Loose Success Score45 | 16 | 4mo ago | ||
| TRIP-Bench Hard PMR | Loose Success36 | 16 | 4mo ago | ||
| TRIP-Bench (Hard FIT) | Loose Success Rate18 | 16 | 4mo ago | ||
| TRIP-Bench Mid | Loose Success Rate55 | 16 | 4mo ago | ||
| TRIP-Bench Easy | DeepSeek-V3.2 | Loose Success Count71 | 16 | 4mo ago | |
| WEBSHOP | ACON | Efficiency Multiplier14 | 15 | 1mo ago | |
| ScienceWorld | AGORA | Efficiency Factor11.5 | 15 | 1mo ago | |
| ALFWORLD | AGORA | Effective Steps Multiplier1 | 15 | 1mo ago | |
| ScienceWorld Unseen | BPO | Average Reward75.67 | 9 | 2mo ago | |
| ScienceWorld Seen | BPO | Average Reward77.1 | 9 | 2mo ago | |
| WebArena (evaluation) | SKILL-DISCO + ReAct | Success Rate (SR)29.1 | 6 | 29d ago | |
| ALFWorld unseen | SKILL-DISCO + CodeAct | Success Rate (SR)99.3 | 6 | 29d ago |