End-to-end motion planning on nuScenes benchmark leaderboard · SOTA2 Research

End-to-end motion planning on nuScenes

0.2L2 Displacement Error (1s)NaviDriveVLM

Evaluation Results

Method	Links
NaviDriveVLM2026.03		0.2	0.5	0.93	0.46	—	—	—	—
Driver-VLM2026.03		0.24	0.65	1.25	0.6	—	—	—	—
DriveVA2026.04		0.33	0.76	1.43	0.84	0	7	0.12	6
DriveVA2026.04		0.33	0.76	1.43	0.84	0	7	0.12	6
GenAD2026.03		0.36	0.83	1.55	0.91	—	—	—	—
Verdi2026.03		0.36	0.62	0.96	0.65	—	—	—	—
GenAD2026.04		0.36	0.83	1.55	0.91	6	23	1	43
VAD-Base2026.03		0.41	0.7	1.05	0.72	—	—	—	—
DriveVLA-W02026.04		0.43	1.26	2.6	1.43	22	66	1.42	77
UniAD2026.03		0.44	0.67	0.96	0.69	—	—	—	—
Ego-MLP2026.03		0.46	0.76	1.12	0.78	—	—	—	—
UniAD2026.04		0.48	0.96	1.65	1.03	5	17	0.71	31
Doe-12026.04		0.5	1.18	2.11	1.26	4	37	1.19	53
OccWorld2026.04		0.52	1.27	2.41	1.4	12	40	2.08	87
VAD-Base2026.04		0.54	1.15	1.98	1.22	4	39	1.17	53
VAD-Tiny2026.04		0.6	1.23	2.06	1.3	31	53	1.33	72
Epona2026.04		0.61	1.17	1.98	1.25	1	22	0.85	36
OccNet2026.04		1.29	2.13	2.99	2.14	21	59	1.37	72
ST-P32026.03		1.33	2.11	2.9	2.11	—	—	—	—
ST-P32026.04		1.33	2.11	2.9	2.11	23	62	1.27	71
OpenEMMA2026.03		1.45	3.21	3.76	2.81	—	—	—	—
PWM2026.04		2.06	3.91	6	3.99	12	15	0.86	36