Long-horizon robot task execution on Chemical Task Suite
0.8714Task ProgressHuman + π sub mem
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Human + π sub memParadigm=Agentic2026.07 | 0.8714 | 75 | |
| CortexParadigm=Agentic2026.07 | 0.7857 | 65 | |
| πmemParadigm=End-to-end2026.07 | 0.2929 | 0 | |
| π0.5Paradigm=End-to-end2026.07 | 0.1786 | 0 |