ResearchBenchmarksOffline Reinforcement Learning on visual-cube-double-play singletask task1 v0Follow27Normalized ScoreGTP20.7622.382425.62Oct 13, 2025Evaluation ResultsMethodMethodLinksNormalized ScoreGTPobservation_type=visua...observation_type=visual-observation, evaluation_seeds=4 random seeds2025.1027FQLobservation_type=visua...observation_type=visual-observation, evaluation_seeds=4 random seeds2025.1021