SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning

Open paper
Benchmark Model Rank Results
robot-manipulation-on-libero-longSimpleVLA-RL (OpenVLA-OFT)#4Success Rate: 98.5
robot-manipulation-on-libero-spatialSimpleVLA-RL (OpenVLA-OFT)#6Success Rate: 99.4
robot-manipulation-on-liberoSimpleVLA-RL (OpenVLA-OFT)#7Success Rate: 99.1
robot-manipulation-on-libero-goalSimpleVLA-RL (OpenVLA-OFT)#9Success Rate: 99.2
robot-manipulation-on-libero-objectSimpleVLA-RL (OpenVLA-OFT)#54Success Rate: 99.1

Rank counts only results with a code link.