ResearchBenchmarksRobotic Control on PointMassSimpleFollow100Success RatioPPO9597.5100102.5Jun 19, 2026Evaluation ResultsMethodMethodLinksSuccess RatioGradient StepsAPG SpeedupTotal StepsPPO2026.061002,150——APG2026.061003206.72—PPO2026.06100——69,120APG2026.06100—1.838,400