ResearchBenchmarksContinuous Control on InvertedPendulum MuJoCo v5Follow1,000Max Evaluation ReturnPPO9509751,0001,025May 1, 2026Evaluation ResultsMethodMethodLinksMax Evaluation ReturnPPOTraining steps=1,000,0...Training steps=1,000,000, Seeds=102026.051,000TD3Training steps=1,000,0...Training steps=1,000,000, Seeds=102026.051,000SACTraining steps=1,000,0...Training steps=1,000,000, Seeds=102026.051,000TQCTraining steps=1,000,0...Training steps=1,000,000, Seeds=102026.051,000SAVGOTraining steps=1,000,0...Training steps=1,000,000, Seeds=102026.051,000