ResearchBenchmarksG1 speed on walker2d medium-replay v2Follow581.5Mean Raw MuJoCo ReturnFrozen34.148176.249318.35460.451Apr 23, 2026Evaluation ResultsMethodMethodLinksMean Raw MuJoCo ReturnFrozen2026.04581.5PoEalpha=0.9alpha=0.92026.04468.3CQL2026.04362.7Additivealpha=0.5alpha=0.52026.04182.2PoEalpha=0.5alpha=0.52026.04171.9KL-Regalpha=0.5alpha=0.52026.04171.9Prior2026.04129.3AWR2026.04109.4IQL2026.0455.2