BridgeVLA++: A Data-Efficient, Generalizable, and Memory-Augmented Vision-Language-Action Framework for 3D Manipulation
Open paper
Benchmark
Model
Rank
Results
robot-manipulation-generalization-on-gembench
BridgeVLA++
#1
Average Success Rate: 51.1
robot-manipulation-on-rlbench
BridgeVLA++
#1
Succ. Rate (18 tasks, 100 demo/task): 93.7
Rank counts only results with a code link.