Multi-step manipulation on DROID Tabletop Multi-step tasks
98Success RateTiPToP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| TiPToPScene=Overall Aggregate, Environment=Mixed2026.03 | 98 | 74.6 | |
| π0.5-DROIDScene=Overall Aggregate, Environment=Mixed2026.03 | 55 | 52.4 | |
| TiPToPScene=Multi-step Aggregate, Environment=Mixed2026.03 | 23 | 75.2 | |
| TiPToPScene=Color cubes → bowl (sim), Environment=Simulation, Evaluation Team=External evaluation team, Number of trials=102026.03 | 9 | 94.6 | |
| π0.5-DROIDScene=Multi-step Aggregate, Environment=Mixed2026.03 | 6 | 52.2 | |
| TiPToPScene=Pack pods → tray†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 4 | 80 | |
| TiPToPScene=Aleve bottle → tray (obs.)†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 4 | 80 | |
| π0.5-DROIDScene=AirPods → cup, Environment=Real-world, Evaluation Team=External evaluation team, Number of trials=52026.03 | 3 | 75 | |
| π0.5-DROIDScene=Aleve bottle → tray (obs.)†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 2 | 70 | |
| TiPToPScene=Three marbles → cup†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 2 | 80 | |
| TiPToPScene=Marbles + cable†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 2 | 70 | |
| TiPToPScene=AirPods → cup, Environment=Real-world, Evaluation Team=External evaluation team, Number of trials=52026.03 | 1 | 55 | |
| π0.5-DROIDScene=Pack pods → tray†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 1 | 65.7 | |
| TiPToPScene=Pack pods → tray (obs.)†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 1 | 67 | |
| π0.5-DROIDScene=Color cubes → bowl (sim), Environment=Simulation, Evaluation Team=External evaluation team, Number of trials=102026.03 | 0 | 24.2 | |
| π0.5-DROIDScene=Pack pods → tray (obs.)†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 0 | 64 | |
| π0.5-DROIDScene=Three marbles → cup†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 0 | 6.7 | |
| π0.5-DROIDScene=Marbles + cable†, Environment=Real-world, Evaluation Team=System designers, Number of trials=52026.03 | 0 | 60 |