ResearchBenchmarksRobotic Manipulation on ManiSkill3 PullCubeMO v1Follow0.41Average RewardSAC0.2020.2560.310.364Mar 5, 2026Evaluation ResultsMethodMethodLinksAverage RewardSACtraining steps=50 [k],...training steps=50 [k], seeds=3, w_target=0.252026.030.41RC-SACtraining steps=50 [k],...training steps=50 [k], seeds=3, w_target=0.252026.030.37TD3training steps=50 [k],...training steps=50 [k], seeds=3, w_target=0.252026.030.26RC-TD3training steps=50 [k],...training steps=50 [k], seeds=3, w_target=0.252026.030.21