Offline Reinforcement Learning on D4RL MuJoCo v2 (test)
77.6HalfCheetah-Medium ScoreRAMBO
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RAMBOMethod Category=Model-based, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 77.6 | 40 | 21.6 | 11.5 | 92.8 | 86.9 | 68.9 | 96.6 | 85 | 93.7 | 83.3 | 68.3 | 826.2 | |
| RAMBO^OFFMethod Category=Model-based, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 71 | 33.5 | 15.5 | 0.2 | 91.2 | 89.1 | 67 | 97.6 | 88.5 | 79.3 | 89.5 | 63.1 | 785.5 | |
| MOPOMethod Category=Model-based, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 69.5 | 35.4 | 4.1 | 4.2 | 48 | -0.2 | 68.2 | 39.1 | 69.4 | 72.7 | 3.3 | -0.3 | 413.4 | |
| SAC-N2026.05 | 68.2 | — | — | — | 40.8 | 87.5 | 60.7 | 100.3 | 79 | 99 | 101.3 | 114.9 | 83.5 | |
| EDAC2026.05 | 67.7 | — | — | — | 101.7 | 93.4 | 62.1 | 99.7 | 87.1 | 104.8 | 105.2 | 114.8 | 92.9 | |
| ReBRAC2026.05 | 64 | — | — | — | 102.3 | 85.8 | 51.2 | 95 | 84.2 | 103.8 | 109.5 | 111.9 | 89.7 | |
| COMBOMethod Category=Model-based, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 54.2 | 38.8 | 17.9 | 7 | 94.9 | 75.5 | 55.1 | 73.1 | 56 | 90 | 111.1 | 96.1 | 769.7 | |
| RepB-SDEMethod Category=Model-based, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 49.1 | 32.9 | 8.6 | 21.1 | 34 | 72.1 | 57.5 | 62.2 | 49.8 | 55.4 | 82.6 | 88.8 | 614.1 | |
| CQLMethod Category=Model-free, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 49 | 19.6 | 6.7 | 2.4 | 66.6 | 83.8 | 47.1 | 97 | 88.2 | 90.8 | 106.8 | 109.4 | 767.4 | |
| UNIQaverages over seeds=0-22026.05 | 48.9 | — | — | — | 75.6 | 85.5 | 46 | 101.6 | 89.4 | 94.8 | 111.8 | 112.9 | 85.2 | |
| TD3+BCMethod Category=Model-free, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 48.3 | 11 | 8.5 | 1.6 | 59.3 | 83.7 | 44.6 | 60.9 | 81.8 | 90.7 | 98 | 110.1 | 698.5 | |
| IQL2026.05 | 48.3 | — | — | — | 67.5 | 80.9 | 44.5 | 97.4 | 82.2 | 94.7 | 107.4 | 111.7 | 81.6 | |
| TD3+BC2026.05 | 48.1 | — | — | — | 60.4 | 82.7 | 44.8 | 64.4 | 85.6 | 90.8 | 101.2 | 110 | 76.4 | |
| IQLMethod Category=Model-free, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 47.4 | — | — | — | 66.3 | 78.3 | 44.2 | 94.7 | 73.9 | 86.7 | 91.5 | 109.6 | 692.6 | |
| CQL2026.05 | 47 | — | — | — | 59.1 | 80.8 | 45 | 95.1 | 73.1 | 95.6 | 99.3 | 109.6 | 78.3 | |
| BC2026.05 | 42.4 | — | — | — | 53.5 | 63.2 | 35.7 | 29.8 | 21.8 | 55.9 | 52.3 | 99 | 50.4 | |
| DT2026.05 | 42.2 | — | — | — | 65.1 | 67.6 | 38.9 | 81.8 | 59.9 | 91.6 | 110.4 | 107.1 | 73.8 | |
| MOReLMethod Category=Model-based, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 42.1 | 25.6 | 53.6 | 37.3 | 95.4 | 77.8 | 40.2 | 93.6 | 49.8 | 53.3 | 108.7 | 95.6 | 773 | |
| BCMethod Category=Model-free, Normalization Procedure=[14], Evaluation Window=last 10 iterations, Seeds=52022.04 | 36.1 | 2.1 | 9.8 | 1.6 | 29 | 6.6 | 38.4 | 11.8 | 11.3 | 35.8 | 111.9 | 6.4 | 300.8 |