Closed-loop Planning on nuPlan 14-hard (test)
86NRLog-replay
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Log-replayType=Expert, Post-processing=True2026.03 | 86 | 65.8 | — | — | — | |
| Log-replayType=Expert2026.01 | 85.96 | 68.8 | — | — | — | |
| Log-replayType=Expert2026.01 | 85.96 | 68.8 | — | — | — | |
| Log-replayType=Expert2025.09 | 85.96 | 68.8 | — | — | — | |
| Log-replayType=Expert2026.02 | 85.96 | 68.8 | — | — | — | |
| CarPLANType=Learning, Post-processing=True2026.03 | 82.2 | 82 | — | — | — | |
| FlowDrive*Type=Rule-based & Hybrid2025.09 | 81.86 | 81.96 | — | — | — | |
| PLUTOType=Learning, Post-processing=True2026.03 | 80.1 | 76.9 | — | — | — | |
| PLUTOType=Rule-based & Hybrid2025.09 | 80.08 | 76.88 | — | — | — | |
| PLUTOType=Rule-based & Hybrid2026.02 | 80.08 | 76.88 | — | — | — | |
| Diffusion-PlannerType=Learning, Post-processing=True2026.03 | 78.9 | 82 | — | — | — | |
| Diffusion PlannerType=Rule-based & Hybrid, Refinement=Enabled2025.09 | 78.87 | 82 | — | — | — | |
| Diffusion Planner w/ refineType=Rule-based & Hybrid2026.02 | 78.87 | 82 | — | — | — | |
| FlowDriveType=Learning-based2025.09 | 77.86 | 73.09 | — | — | — | |
| PlannerRFTType=Learning2026.01 | 77.16 | 72.21 | — | — | — | |
| PlannerRFTType=Learning2026.01 | 77.16 | 72.21 | — | — | — | |
| FlowDrive-Type=Learning-based2025.09 | 77.07 | 69.46 | — | — | — | |
| RaScType=Learning, Post-processing=True2026.03 | 77 | 65 | — | — | — | |
| Flow PlannerType=Learning2026.01 | 76.47 | 70.42 | — | — | — | |
| Flow PlannerType=Learning2026.01 | 76.47 | 70.42 | — | — | — | |
| RAPiDType=Learning-based2026.02 | 76.09 | 66.94 | — | — | — | |
| Diffusion PlannerDDIMType=Learning, Sampler=5-step DDIM2026.01 | 76.01 | 68.18 | — | — | — | |
| Diffusion PlannerType=Learning2026.01 | 75.99 | 69.22 | — | — | — | |
| Diffusion PlannerDPMType=Learning, Sampler=10-step DPM-solver2026.01 | 75.99 | 69.22 | — | — | — | |
| Diffusion PlannerType=Learning-based2025.09 | 75.99 | 69.22 | — | — | — | |
| DiffusionPlannerType=Learning-based2026.02 | 75.99 | 69.22 | — | — | — | |
| PLUTOType=Learning2026.01 | 70.03 | 59.74 | — | — | — | |
| PLUTOType=Learning2026.01 | 70.03 | 59.74 | — | — | — | |
| PLUTOType=Learning-based, Refinement=None2025.09 | 70.03 | 59.74 | — | — | — | |
| PLUTO w/o refineType=Learning-based, Pre-computed reference paths=true2026.02 | 70.03 | 59.74 | — | — | — | |
| PlanTFType=Learning2026.01 | 69.7 | 61.61 | — | — | — | |
| PlanTFType=Learning2026.01 | 69.7 | 61.61 | — | — | — | |
| PlanTFType=Learning-based2025.09 | 69.7 | 61.61 | — | — | — | |
| PlanTFType=Learning-based2026.02 | 69.7 | 61.61 | — | — | — | |
| GameFormerType=Rule-based & Hybrid2025.09 | 68.7 | 67.05 | — | — | — | |
| GameFormerType=Rule-based & Hybrid2026.02 | 68.7 | 67.05 | — | — | — | |
| GameFormerType=Learning, Post-processing=True2026.03 | 68.7 | 67.1 | — | — | — | |
| PDM-ClosedType=Rule, Post-processing=True2026.03 | 66 | 76.1 | — | — | — | |
| PDM-HybridType=Learning, Post-processing=True2026.03 | 66 | 76.1 | — | — | — | |
| PDM-HybridType=Rule-based & Hybrid2025.09 | 65.99 | 76.07 | — | — | — | |
| PDM-HybridType=Rule-based & Hybrid2026.02 | 65.99 | 76.07 | — | — | — | |
| PDM-ClosedType=Rule2026.01 | 65.08 | 75.19 | — | — | — | |
| PDM-ClosedType=Rule2026.01 | 65.08 | 75.19 | — | — | — | |
| PDM-ClosedType=Rule-based & Hybrid2025.09 | 65.08 | 75.19 | — | — | — | |
| PDM-ClosedType=Rule-based & Hybrid2026.02 | 65.08 | 75.19 | — | — | — | |
| ReflexDiffusionType=Learning-based2026.01 | 59.94 | 65.53 | — | — | — | |
| Diffusion PlannerType=Learning-based2026.01 | 58.47 | 57.41 | — | — | — | |
| IDMType=Rule, Post-processing=True2026.03 | 56.2 | 62.3 | — | — | — | |
| IDMType=Rule2026.01 | 56.15 | 62.26 | — | — | — | |
| IDMType=Rule2026.01 | 56.15 | 62.26 | — | — | — | |
| IDMType=Rule-based & Hybrid2025.09 | 56.15 | 62.26 | — | — | — | |
| IDMType=Rule-based & Hybrid2026.02 | 56.15 | 62.26 | — | — | — | |
| GameformerType=Hybrid2026.01 | 53.12 | 57.46 | — | — | — | |
| UrbanDriverType=Learning-based2026.02 | 50.4 | 49.95 | — | — | — | |
| Diffusion-esType=Learning-based2026.01 | 44.63 | 52.72 | — | — | — | |
| Diffusion PlannerType=Learning-based, setting=conditional dropout+cfg2026.01 | 44.4 | 55.55 | — | — | — | |
| SAH-DriveType=Hybrid2026.01 | 43.08 | 57.4 | — | — | — | |
| PlutoType=Learning-based2026.01 | 42.21 | 45.98 | — | — | — | |
| PlanTFType=Learning-based2026.01 | 38.24 | 42.52 | — | — | — | |
| IDMType=Rule-based2026.01 | 36.74 | 62.42 | — | — | — | |
| PDM-OpenType=Learning2026.01 | 33.51 | 35.83 | — | — | — | |
| PDM-OpenType=Learning2026.01 | 33.51 | 35.83 | — | — | — | |
| PDM-OpenType=Learning-based2025.09 | 33.51 | 33.53 | — | — | — | |
| PDM-OpenType=Learning-based, Pre-computed reference paths=true2026.02 | 33.51 | 35.83 | — | — | — | |
| PDM-ClosedType=Rule-based2026.01 | 32.55 | 53.03 | — | — | — | |
| PDM-HybridType=Hybrid2026.01 | 32.54 | 53.04 | — | — | — | |
| UrbanDriverType=Learning-based2026.01 | 26.09 | 14.66 | — | — | — | |
| PlanCNNType=Learning-based2026.01 | 20.16 | 28.03 | — | — | — | |
| Diffusion PlannerType=Learning-based, setting=conditional dropout2026.01 | 12.6 | 14.04 | — | — | — | |
| GameFormerType=Learning2026.01 | 7.08 | 6.69 | — | — | — | |
| GameFormerType=Learning2026.01 | 7.08 | 6.69 | — | — | — | |
| GameFormerType=Learning-based, Refinement=None2025.09 | 7.08 | 6.69 | — | — | — | |
| GameFormer w/o refineType=Learning-based2026.02 | 7.08 | 6.69 | — | — | — | |
| DFPTraining Paradigm=Generative IL, Simulation Protocol=NR2026.06 | — | — | — | 76.91 | — | |
| DFPTraining Paradigm=Generative IL, Simulation Protocol=R2026.06 | — | — | — | 63.56 | — | |
| DFP-FMTraining Paradigm=Generative IL, Simulation Protocol=NR2026.06 | — | — | — | 79.43 | — | |
| DFP-FMTraining Paradigm=Generative IL, Simulation Protocol=R2026.06 | — | — | — | 67.94 | — | |
| Diffusion PlannerType=Rule-based & Hybrid, Refinement=true2025.11 | — | — | 78.91 | — | — | |
| Diffusion PlannerType=Learning-based2025.11 | — | — | 75.44 | — | — | |
| Diffusion PlannerType=Rule-based & Hybrid, Refinement=true2025.11 | — | — | 81.66 | — | — | |
| Diffusion PlannerType=Learning-based2025.11 | — | — | 68.95 | — | — | |
| Diffusion PlannerReactive setting=Non-reactive (NR), Refinement module=with refinement2026.04 | — | — | — | — | 78.87 | |
| Diffusion PlannerReactive setting=Reactive (R), Refinement module=with refinement2026.04 | — | — | — | — | 82 | |
| Diffusion PlannerReactive setting=Non-reactive (NR)2026.04 | — | — | — | — | 75.67 | |
| Diffusion PlannerReactive setting=Reactive (R)2026.04 | — | — | — | — | 68.56 | |
| Diffusion PlannerTraining Paradigm=Generative IL, Simulation Protocol=NR2026.06 | — | — | — | 75.99 | — | |
| Diffusion PlannerTraining Paradigm=Generative IL, Simulation Protocol=R2026.06 | — | — | — | 69.22 | — | |
| DTPPCategory=Data-driven2025.05 | — | — | — | 59.44 | — | |
| ExpertTraining Paradigm=Human log replay, Simulation Protocol=NR2026.06 | — | — | — | 85.96 | — | |
| ExpertTraining Paradigm=Human log replay, Simulation Protocol=R2026.06 | — | — | — | 68.8 | — | |
| Flow PlannerTraining Paradigm=Generative IL, Simulation Protocol=NR2026.06 | — | — | — | 76.47 | — | |
| Flow PlannerTraining Paradigm=Generative IL, Simulation Protocol=R2026.06 | — | — | — | 70.42 | — | |
| GameFormerCategory=Data-driven2025.05 | — | — | — | 66.59 | — | |
| GameFormerReactive setting=Non-reactive (NR)2026.04 | — | — | — | — | 68.7 | |
| GameFormerReactive setting=Reactive (R)2026.04 | — | — | — | — | 67.05 | |
| GC-PGPType=Learning-based2025.11 | — | — | 46.13 | — | — | |
| GC-PGPType=Learning-based2025.11 | — | — | 42.74 | — | — | |
| GC-PGPCategory=Data-driven2025.05 | — | — | — | 43.22 | — | |
| IDMType=Rule-based & Hybrid2025.11 | — | — | 56.15 | — | — | |
| IDMType=Rule-based & Hybrid2025.11 | — | — | 62.26 | — | — |