Offline Reinforcement Learning on D4RL Antmaze (All Configurations)
89.2Medium Diverse Success RateDfsrLite
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| DfsrLite2025.10 | 89.2 | — | — | 88 | 72.4 | 80.4 | 82.5 | — | — | |
| HD2025.10 | 88.7 | — | — | — | — | 83.6 | — | — | — | |
| DV2025.10 | 87.4 | — | — | 89 | 76.4 | 80 | 83.2 | — | — | |
| MDD2025.10 | 87.3 | — | — | 88.7 | 77.3 | 80.7 | 83.5 | — | — | |
| QIPO-OTPolicy Category=Diffusion, Training Steps=1M, Batch Size=2562026.05 | 86.4 | 93.6 | 76.1 | 80 | 55.5 | 32.1 | 70.6 | — | — | |
| QIPO-DiffPolicy Category=Diffusion, Training Steps=1M, Batch Size=2562026.05 | 86 | 97.5 | 73.9 | 82.8 | 73.3 | 40.5 | 75.7 | — | — | |
| FACArchitecture Type=MLP based2026.02 | 85 | 98.5 | 93.5 | 88 | 90 | 88 | 90.5 | — | — | |
| IDQL2025.10 | 84.8 | — | — | 84.5 | 63.5 | 67.9 | 75.2 | — | — | |
| QGPOPolicy Category=Diffusion, Training Steps=1M, Batch Size=2562026.05 | 83.8 | 96.4 | 74.4 | 83.6 | 66.6 | 64.8 | 78.3 | — | — | |
| IDQL*2025.10 | 83.3 | — | — | 67.3 | 48.7 | 40 | 69.8 | — | — | |
| DQL*2025.10 | 82.6 | — | — | 87.3 | 81.3 | 70.6 | 80.5 | — | — | |
| SfBC2025.10 | 82 | — | — | 81.3 | 59.3 | 45.5 | 67 | — | — | |
| DQL2025.10 | 78.6 | — | — | 76.6 | 46.4 | 56.6 | 64.6 | — | — | |
| IFQLPolicy Category=Flow, Training Steps=1M, Batch Size=2562026.05 | 72.3 | 84.3 | 74.5 | 59.9 | 49.9 | 55.5 | 66.1 | — | — | |
| Q-FlowPolicy Category=Flow, Training Steps=1M, Batch Size=2562026.05 | 71.8 | 95 | 88.8 | 77.7 | 73.3 | 75.9 | 80.4 | — | — | |
| FBRACPolicy Category=Flow, Training Steps=1M, Batch Size=2562026.05 | 54.3 | 92.8 | 64.2 | 67.5 | 30.8 | 31 | 56.8 | — | — | |
| FQLPolicy Category=Flow, Training Steps=1M, Batch Size=2562026.05 | 54 | 96 | 88.8 | 73.6 | 69.3 | 75.7 | 76.2 | — | — | |
| FAWACPolicy Category=Flow, Training Steps=1M, Batch Size=2562026.05 | 45 | 89.9 | 60.9 | 49 | 9.3 | 13.2 | 44.6 | — | — | |
| ROMIseeds=52026.03 | 30.1 | 70.3 | 32.8 | 51.3 | 0 | 6.4 | — | 186.5 | — | |
| RAMBOseeds=52026.03 | 20.4 | 42.6 | 12.9 | 18.7 | 0 | 0 | — | 94.6 | — | |
| ReinformerArchitecture Type=Transformer based2026.02 | 10.6 | 84.4 | 65.8 | 13.2 | — | — | — | — | — | |
| DD2025.10 | 4 | — | — | 8 | 0 | 0 | 3 | — | — | |
| Diffuser2025.10 | 2 | — | — | 6.7 | 17.3 | 27.3 | 13.3 | — | — | |
| MOBILEseeds=52026.03 | 1.6 | 77 | 20.4 | 64.6 | 2.6 | 7.2 | — | 173.4 | — | |
| MOPOseeds=52026.03 | 0 | 0 | 0 | 0 | 0 | 0 | — | 0 | — | |
| COMBOseeds=52026.03 | 0 | 80.3 | 57.3 | 0 | 0 | 0 | — | 137.6 | — | |
| BC2025.10 | 0 | — | — | 0 | 0 | 0 | 0 | — | — | |
| CACPolicy Category=Distillation Policies2026.05 | — | — | — | — | — | — | — | 30 | 26 | |
| DCArchitecture Type=Transformer based2026.02 | — | 85 | 78.5 | — | — | — | — | — | — | |
| DMArchitecture Type=SSM based2026.02 | — | 100 | 90 | — | — | — | — | — | — | |
| DSRLPolicy Category=Latent Opt.2026.05 | — | — | — | — | — | — | — | 56 | 40 | |
| FAVPolicy Category=Ours2026.05 | — | — | — | — | — | — | — | 80 | 54 | |
| FAV-AdaptivePolicy Category=Ours2026.05 | — | — | — | — | — | — | — | 79 | 50 | |
| FAWACPolicy Category=Flow-based Policies2026.05 | — | — | — | — | — | — | — | 44 | 19 | |
| FQLPolicy Category=Distillation Policies2026.05 | — | — | — | — | — | — | — | 84 | 48 | |
| IFQLPolicy Category=Flow-based Policies2026.05 | — | — | — | — | — | — | — | 65 | 34 | |
| IQLPolicy Category=Gaussian Policies2026.05 | — | — | — | — | — | — | — | 57 | 27 | |
| QAMPolicy Category=Flow-based Policies2026.05 | — | — | — | — | — | — | — | 80 | 45 | |
| ReBRACPolicy Category=Gaussian Policies2026.05 | — | — | — | — | — | — | — | 78 | 36 | |
| SRPOPolicy Category=Distillation Policies2026.05 | — | — | — | — | — | — | — | 74 | 20 |