Open-loop Trajectory Prediction on nuScenes v1.0 (test) benchmark leaderboard · SOTA2 Research

Open-loop Trajectory Prediction on nuScenes v1.0 (test)

0.13L2 Error (1s)EMMA+

Evaluation Results

Method
EMMA+2026.01	0.13	0.27	0.48	0.29
ImpromptuVLA(3B)2026.01	0.13	0.27	0.52	0.3
ImpromptuVLA(7B)2026.01	0.13	0.27	0.53	0.3
LatentVLA(iPad)2026.01	0.13	0.28	0.56	0.33
OmniDrive2026.01	0.14	0.29	0.55	0.33
EMMA2026.01	0.14	0.29	0.54	0.32
LatentVLA(Transfuser)2026.01	0.14	0.29	0.58	0.34
Distilled LatentVLA(iPad)2026.01	0.14	0.3	0.6	0.35
Ego-MLP*2026.01	0.15	0.32	0.59	0.35
DriveVLM-Dual2026.01	0.15	0.29	0.48	0.31
EMMA (random init)2026.01	0.15	0.33	0.63	0.37
Distilled LatentVLA(Transfuser)2026.01	0.15	0.31	0.62	0.36
BEV-Planner2026.01	0.16	0.32	0.57	0.35
VAD2026.01	0.17	0.34	0.6	0.37
DriveVLM2026.01	0.18	0.34	0.68	0.4
GPT-4o2026.01	0.28	0.93	2.02	1.07
Claude-3.7-Sonnet2026.01	0.28	0.94	2.04	1.09
Claude-3.5-Sonnet2026.01	0.29	0.98	2.12	1.13
Gemini-2.0-Flash2026.01	0.31	1.08	2.36	1.25
Gemini-2.5-Pro2026.01	0.37	1.35	2.96	1.56
DeepSeek-VL2-28B2026.01	0.37	1.35	2.96	1.56
UniAD2026.01	0.42	0.64	0.91	0.66
Qwen-2.5-VL-7B-Instruct2026.01	0.46	1.33	2.55	1.45
LLaMA-3.2-11B-Vision-Instruct (superscript 1)2026.01	0.52	1.42	2.68	1.54
DeepSeek-VL2-16B2026.01	0.66	1.68	2.92	1.75
LLaMA-3.2-90B-Vision-Instruct2026.01	0.66	1.71	3.01	1.79
Qwen2-VL-7B-Instruct2026.01	1.45	3.21	3.76	2.81
LLaVA-1.6-Mistral-7B2026.01	1.49	3.38	4.09	2.98
Llama-3.2-11B-Vision-Instruct (superscript 2)2026.01	1.54	3.31	3.91	2.92