ReconVLA: Reconstructive Vision-Language-Action Model as Effective Robot Perceiver

Open paper
Benchmark Model Rank Results
robot-manipulation-on-calvin-abcd-to-dReconVLA (CALVIN ABCD→D)#9Average Length: 4.23
robot-manipulation-on-calvinReconVLA–avg. sequence length (D to D): 3.95

Rank counts only results with a code link.