Offline Reinforcement Learning on D4RL MuJoCo Locomotion Domain v2 benchmark leaderboard · SOTA2 Research

Offline Reinforcement Learning on D4RL MuJoCo Locomotion Domain v2

96Return (HalfCheetah, M-E)SSAR+PAR

Evaluation Results

Method	Links
SSAR+PAR2026.02		96	112.7	111.7	58.5	102.5	89.3	49.7	102.6	95.1	90.9
PG2026.02		95.2	110.4	109.4	45.6	97.5	82.3	46.4	91.3	83.7	81.1
TD3+BC+PAR2026.02		95.2	111.8	111.5	49.4	66.7	85.8	45.5	95.8	87.3	83.2
QIPO2026.02		94.5	108	110.9	54.2	94.1	97.3	48	101.3	78.6	87.9
EDP+PAR2026.02		93.3	110.9	111.9	54.9	86.4	89.5	48.3	101.7	94.9	88.6
IQL+PAR2026.02		93.2	108.3	111.2	48.4	71.3	82.5	44.5	95.7	82.8	82
RefPlan2026.02		92.8	57.8	114	74.6	32.8	91.6	76.3	82.6	91.2	78.8
TAT2026.02		92.5	109.4	108.8	44.3	82.6	81	39.2	95.3	78.2	78.9
LoMAP2026.02		91.1	110.6	109.2	45.4	93.7	79.9	39.1	97.6	78.7	82.9
FQL2026.02		90.1	86.2	100.5	60.1	74.5	72.7	51.1	85.4	82.1	78