ResearchBenchmarksOffline Goal-Conditioned Reinforcement Learning on puzzle 4x6-play v0Follow18Average Binary Success RateHIQL2vr-0.724.14913.86May 21, 2026Evaluation ResultsMethodMethodLinksAverage Binary Success RateHIQL2vrSize=5M, Dim.=115Size=5M, Dim.=1152026.0518ARLiSize=5M, Dim.=115Size=5M, Dim.=1152026.0514IQLSize=5M, Dim.=115Size=5M, Dim.=1152026.0513ARLeSize=5M, Dim.=115Size=5M, Dim.=1152026.059HIQL1vrSize=5M, Dim.=115Size=5M, Dim.=1152026.050HIQL2vSize=5M, Dim.=115Size=5M, Dim.=1152026.050