Robo-bench leaderboard

Robo-bench (Vals AI Robo-bench) is an external benchmark published by Robo-bench leaderboard and methodology. Tests agents controlling a simulated robot arm and humanoid through motor commands, sensor readings, and camera images. Its evaluation set is described by the source as 24 simulated robot-control tasks. The source rates its difficulty as real-time perception, planning, and motor control.

Each model's result is published by Robo-bench leaderboard and methodology as mean task score out of 100. ModelScale prints that value to 2 decimal places, exactly as published, and never converts or rescales it. This page ranks the 0 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.