ResearchBenchmarksMulti-Objective Reinforcement Learning on mo-hopper v5Multi-Objective Reinforcement Learning on mo-hopper v5Follow15,800,000Hypervolume (x10^7)PRISM8,104,00010,102,00012,100,00014,098,000Feb 20, 2026Evaluation resultsMethodMethodLinksHypervolume (x10^7)EUMVOPRISMReward Sparsity=Extrem...Reward Sparsity=Extreme (first objective)2026.0215,800,000147.4366.66OracleReward Sparsity=None (...Reward Sparsity=None (Dense)2026.0213,000,000129.0459.07BaselineReward Sparsity=Extrem...Reward Sparsity=Extreme (first objective)2026.028,400,00097.6443.36