Towards Generalizable Vision-Language Robotic Manipulation: A Benchmark and LLM-guided 3D Policy

Benchmark Model Rank Results
robot-manipulation-generalization-on-gembench3D-LOTUS++#2Average Success Rate: 48.0Average Success Rate (L1): 68.7±0.6…
robot-manipulation-generalization-on-gembench3D-LOTUS#3Average Success Rate: 45.7Average Success Rate (L1): 94.3±1.4…
robot-manipulation-on-rlbench3D-LOTUS#5Succ. Rate (18 tasks, 100 demo/task): 83.1…