ResearchBenchmarksTable Pick Place on Real-world manipulation dataset R1 humanoid (TPP-3)Follow36Success RateViTaS23.5226.763033.24Feb 12, 2026Evaluation ResultsMethodMethodLinksSuccess RateViTaSInput=head camera and...Input=head camera and tactile sensor2026.0236DPInput=head, left, righ...Input=head, left, right cameras, Encoder=learning-from-scratch CNN encoder2026.0224