| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| State Abstraction | AMA-Bench SA | Success Metric52.92 | 20 | |
| State Updating | AMA-Bench (SU) | Success Metric58.73 | 20 | |
| Causal Inference | AMA-Bench (CI) | Success Metric57.21 | 20 | |
| Agent Memory | AMA-Bench real-world | Recall (Accuracy)62.38 | 14 | |
| Embodied AI | AMA-bench Embodied AI Domain | TTFT Mean (std) [s]0.137 | 8 | |
| Trajectory QA | AMA-Bench Full 208-episode trajectory | F1 Score38.2 | 5 |