ResearchBenchmarksSafe Reinforcement Learning on SafeAntMaze CshapeFollow88Final Success RateITES-3.5220.244467.76Jun 21, 2026Evaluation ResultsMethodMethodLinksFinal Success RateFinal CostITES2026.068815.2TD3L/R2026.063883.6PPOLAG2026.062041.2TD3LAG2026.0613133.45SafeDreamer2026.06013CUP2026.0606.8FOCOPS2026.06019.4