Embodied Video Model Takes the Lead on RBench

HeyToha · x · 2026-07-12

The post shared additional evaluation results from the paper: the model achieved an average score of 0.620 on RBench, taking the overall lead among the models listed.

It performs particularly well in the following tasks:

This indicates that the approach focuses not just on video quality but also closely aligns with the demands of robotic tasks.

Related event: LingBot-VA/VLA 2.0 Released: Native Embodied Foundation Model(24 posts)→

Original post →

More from Embodied

Embodied channel →