ResearchBenchmarksOffline Goal-Conditioned Reinforcement Learning on cube-double-play v0Follow67Average Binary Success RateARLe-2.6815.4133.551.59May 21, 2026Evaluation ResultsMethodMethodLinksAverage Binary Success RateARLeSize=1M, Dim.=37Size=1M, Dim.=372026.0567ARLiSize=1M, Dim.=37Size=1M, Dim.=372026.0553IQLSize=1M, Dim.=37Size=1M, Dim.=372026.0550HIQL2vrSize=1M, Dim.=37Size=1M, Dim.=372026.053HIQL1vrSize=1M, Dim.=37Size=1M, Dim.=372026.052HIQL2vSize=1M, Dim.=37Size=1M, Dim.=372026.050