Closed-loop Planning on nuPlan (val14)
93.7CLS-NRLog Replay
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Log ReplayType=Expert2024.09 | 93.7 | 98.7 | 94.4 | 98.1 | 99.3 | 98.9 | 96.5 | 81.2 | — | |
| Log-replay (Expert)2026.07 | 93.53 | — | — | — | — | — | — | 80.32 | — | |
| PDM-ClosedType=Rule2024.09 | 93.2 | 98.1 | 93.3 | 99.8 | 95.5 | 92.1 | 99.8 | 93 | — | |
| Intent2Drive2026.07 | 93.02 | — | — | — | — | — | — | 84.05 | — | |
| LLM-ASSIST2026.07 | 93 | — | — | — | — | — | — | 92.2 | — | |
| PLUTO2026.07 | 92.88 | — | — | — | — | — | — | 76.88 | — | |
| PDM-Closed2026.07 | 92.84 | — | — | — | — | — | — | 92.12 | — | |
| PDM-Hybrid2026.07 | 92.77 | — | — | — | — | — | — | 92.11 | — | |
| Flow Planner2026.07 | 90.43 | — | — | — | — | — | — | 83.31 | — | |
| Diffusion Planner2026.07 | 89.87 | — | — | — | — | — | — | 82.8 | — | |
| PLUTOType=Learning2024.09 | 89 | 96.1 | 93.3 | 98.5 | 96.4 | 89.5 | 98.1 | 80 | — | |
| BeTopNetType=Learning2024.09 | 88.3 | 96.6 | 91.6 | 99.5 | 93.2 | 86.6 | 97.1 | 83.7 | — | |
| KnowDiffuserType=Learning-based2026.03 | 87.5 | — | — | — | — | — | — | 81.25 | — | |
| PlanTFType=Learning2024.09 | 85.3 | 94.1 | 90.7 | 96.8 | 93.7 | 89.8 | 97.7 | 77.1 | — | |
| PlanTFType=Learning-based2026.03 | 84.83 | — | — | — | — | — | — | 76.78 | — | |
| PlanTF2026.07 | 84.27 | — | — | — | — | — | — | 76.95 | — | |
| GameFormerType=Hybrid2024.09 | 82.9 | 94.3 | 86.7 | 94.8 | 93.3 | 89 | 98.7 | 83.8 | — | |
| GameFormerType=Learning-based2026.03 | 80.8 | — | — | — | — | — | — | 79.31 | — | |
| GameFormer2026.07 | 79.94 | — | — | — | — | — | — | 79.78 | — | |
| IDMType=Rule2024.09 | 79.3 | 90.9 | 83.4 | 94.1 | 94.4 | 86.2 | 97.3 | 79.3 | — | |
| IDMType=Rule-based2026.03 | 77.02 | — | — | — | — | — | — | 76.5 | — | |
| IDM2026.07 | 75.6 | — | — | — | — | — | — | 77.33 | — | |
| PlanCNNType=Learning-based2026.03 | 72.89 | — | — | — | — | — | — | 71.75 | — | |
| UrbanDriver2026.07 | 68.57 | — | — | — | — | — | — | 64.11 | — | |
| UrbanDriverType=Learning2024.09 | 67.7 | 85.6 | 80.3 | 90.8 | 100 | 80.8 | 91.5 | 64.8 | — | |
| PlanCNNType=Learning2024.09 | 66.9 | 86.9 | 81.4 | 85 | 81.4 | 80.6 | 98 | 64.6 | — | |
| UrbanDriverType=Learning-based2026.03 | 63.27 | — | — | — | — | — | — | 61.02 | — | |
| GC-PGPType=Learning2024.09 | 61.1 | 85.8 | 80.1 | 89.7 | 90 | 60.3 | 99.3 | 54.9 | — | |
| GC PGPType=Learning-based2026.03 | 55.99 | — | — | — | — | — | — | 51.39 | — | |
| PDM-OpenType=Learning2024.09 | 50.2 | 74.5 | 69.1 | 87.9 | 99.5 | 69.8 | 97.7 | 54.8 | — | |
| CoPlannerType=Learning-based, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 89.48 | |
| CoPlannerType=Learning-based, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 79 | |
| DFPType=Learning-based, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 90.33 | |
| DFPType=Learning-based, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 79.97 | |
| DFP-FMType=Learning-based, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 92.68 | |
| DFP-FMType=Learning-based, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 81.3 | |
| Diffusion PlannerType=Learning-based, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 89.87 | |
| Diffusion PlannerType=Learning-based, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 82.8 | |
| Diffusion Planner*Type=Learning-based, Evaluation Protocol=NR, Note=best reimplementation result2026.06 | — | — | — | — | — | — | — | — | 87.87 | |
| Diffusion Planner*Type=Learning-based, Evaluation Protocol=R, Note=best reimplementation result2026.06 | — | — | — | — | — | — | — | — | 77.48 | |
| GameFormerType=Learning-based, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 13.32 | |
| GameFormerType=Learning-based, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 8.69 | |
| Log-replayType=Expert, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 93.53 | |
| Log-replayType=Expert, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 80.32 | |
| PDM-OpenType=Learning-based, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 53.53 | |
| PDM-OpenType=Learning-based, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 54.24 | |
| PlanTFType=Learning-based, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 84.27 | |
| PlanTFType=Learning-based, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 76.95 | |
| PLUTOType=Learning-based, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 88.89 | |
| PLUTOType=Learning-based, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 78.11 | |
| UrbanDriverType=Learning-based, Evaluation Protocol=NR2026.06 | — | — | — | — | — | — | — | — | 68.57 | |
| UrbanDriverType=Learning-based, Evaluation Protocol=R2026.06 | — | — | — | — | — | — | — | — | 64.11 |