Offline Reinforcement Learning on D4RL Gym Locomotion v2 (All Conditions) benchmark leaderboard · SOTA2 Research

Offline Reinforcement Learning on D4RL Gym Locomotion v2 (All Conditions)

73.1HalfCheetah Medium ScoreMOPO

Evaluation Results

Method	Links
MOPO2024.10		73.1	38.3	41.2	69.2	32.7	73.7	70.3	60.6	77.4	35.9	16.7	4.2	593.3
OneStep2024.10		50.4	87.5	84.8	42.7	98.5	61.7	75.1	108.6	111.3	2.3	5.6	6.9	735.4
TD3BC2024.10		48.3	59.3	83.7	44.6	60.9	81.8	90.7	98	110.1	11	8.5	1.6	698.5
IQL2024.10		47.4	66.2	78.3	44.2	94.7	73.8	86.7	91.5	109.6	13.1	7.9	5.4	718.8
CQL2024.10		47	53	73.3	45.5	88.7	81.8	75.6	105.6	107.9	17.5	7.9	5.1	708.9
TT2025.10		46.9	61.1	79	41.9	91.5	82.6	95	110	101.9	—	—	—	78.9
SCAS2024.10		46.6	102.5	82.3	44	101.6	78.1	91.7	109.7	108.4	12.2	31.4	1.4	810.1
SDC2024.10		45.9	64.7	82.7	45.1	94.8	78.5	76.3	99.9	109.2	14.2	3.1	0.2	714.6
OSR2024.10		45.1	62	80.1	43.3	42.1	78.1	63.7	78.9	108.1	1.6	3.7	-0.1	606.7
CQL2025.10		44.4	58	72.5	45.5	95	77.2	91.6	105.4	108.8	—	—	—	77.6
RATE2025.10		43.5	77.4	80.7	39	83.7	73.7	87.4	112.5	108.7	—	—	—	78.5
ELMUR2025.10		43.2	78	79.4	40.4	64.8	75	86.4	111.7	108.9	—	—	—	76.4
BC2025.10		43.1	63.9	77.3	4.3	27.6	36.9	59.9	79.6	36.6	—	—	—	47.7
ELMUR2025.10		43.1	78.6	79.5	39.8	87.6	60.8	89.1	85.3	108.4	—	—	—	74.7
DT2025.10		42.6	67.6	74	36.6	82.7	66.6	86.8	107.6	108.1	—	—	—	74.7
BC2024.10		42	56.2	71	36.4	21.8	24.9	59.6	51.7	101.2	2.6	4.1	1.2	472.7