ResearchBenchmarksContinuous Control on Cartpole (test)Continuous Control on Cartpole (test)Follow478.6Best Test RewardSAC-21.328108.461238.25368.039Mar 2, 2016Nov 20, 2017Aug 11, 2019May 1, 2021Jan 19, 2023Oct 9, 2024Jun 30, 2026Evaluation resultsMethodMethodLinksBest Test RewardEpisodes to Near-OptimalViolationsSAC2026.06478.6—74.4AutoSafe2026.06477.7—0Residual2026.06477.6—16.4Lyapunov2026.06475.2—19Lagrangian2026.06474.9—99.2AdaLam2026.06473.7—39.8CBF2026.06448.2—43.8SimplexRL2026.06384.5—0DDPG2016.03-0.601420—NAF2016.03-0.604190—Random Agent2016.03-2.1——