Safe Reinforcement Learning on Safety Gymnasium Ant (C=25) benchmark leaderboard · SOTA2 Research

Safe Reinforcement Learning on Safety Gymnasium Ant (C=25)

3,231.6ReturnCSPO

Evaluation Results

Method	Links
CSPO2026.06		3,231.6	24.5
CPPOPID2026.06		3,190	25.3
PPO-Lag2026.06		3,134.17	24.3
IPO2026.06		3,123.6	26.6
CPO2026.06		3,055.8	19.6
C-TRPO2026.06		3,042.64	18
APPO2026.06		2,961.36	24.49
FOCOPS2026.06		2,860.88	26.7
CUP2026.06		2,839.65	26.6
EPO2026.06		2,754.55	13.9
P3O2026.06		2,514.81	23.1
PCPO2026.06		1,728.74	18.5