Multi-Robot Collaboration on RoCoBench Arrange Cabinet
100Success RateREFLEX + GPT-4
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| REFLEX + GPT-4Framework=REFLEX, Base LLM=GPT-42025.05 | 100 | 4 | 0 | |
| REFLEX + LLaMA-3.1Framework=REFLEX, Base LLM=LLaMA-3.12025.05 | 95 | 4 | 1.7 | |
| Central PlanFramework=Central Plan2025.05 | 90 | 4 | 2.7 | |
| RoCo + GPT-4Framework=RoCo, Base LLM=GPT-42025.05 | 75 | 4.7 | 2 |