ResearchBenchmarksRun on Humanoid-benchFollow2,056.9Maximum ReturnIRM-63.244487.1781,037.61,588.022Mar 9, 2026Evaluation ResultsMethodMethodLinksMaximum ReturnConvergence StepsIRM2026.032,056.91DreamerV32026.03298.56TD-MPC22026.0366.12PPO2026.0328.85.3SAC2026.0318.35.4