Trajectory Planning on NAVSIM
98.3NC ScoreMiMo-Embodied+RL
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MiMo-Embodied+RLParams=7B, Input=C, Token=796, mode=Reinforcement Learning (RL)2025.11 | 98.3 | 98.1 | 95.5 | 100 | 86.3 | 91 | |
| MiMo-Embodied-ILParams=7B, Input=C, Token=796, mode=Imitation Learning (IL)2025.11 | 98.2 | 94.7 | 94.1 | 100 | 82 | 87.4 | |
| ReCogDrive-Large-ILParams=8B, Input=C, Token=2304, mode=Imitation Learning (IL)2025.11 | 98.1 | 94.5 | 94.2 | 100 | 80.9 | 86.5 | |
| MiMo-EmbodiedParams=7B, Input=C, Token=7962025.11 | 97.9 | 94.3 | 93.8 | 100 | 81.7 | 86.5 | |
| ReCogDrive-Large-RLParams=8B, Input=C, Token=2304, mode=Reinforcement Learning (RL)2025.11 | 97.9 | 97.3 | 94.9 | 100 | 86.9 | 90.4 | |
| InternVL3Params=8B, Input=C, Token=40962025.11 | 97.4 | 93.7 | 93.2 | 100 | 81.2 | 86 |