Our new X account is live! Follow @wizwand_team for updates

Multimodal Autonomous Driving Capability on DriveBench

45.63Perception Accuracy

Qwen2.5-VL-3B-Instruct (Multi)

Updated 4d ago

Evaluation Results

Method	Links
Qwen2.5-VL-3B-Instruct (Multi) 2026.01		45.63	36.41	54.07	45.37
Qwen2.5-VL-3B-Instruct (Single) 2026.01		45.59	34.47	51.97	44.01
GPT-5.1 2026.01		45.56	41.02	51.94	46.17
InternVL3-1B (Multi) 2026.01		43.5	33.15	55.51	44.05
InternVL3-1B (Single) 2026.01		43.13	34.32	52.97	43.47
InternVL3-78B 2026.01		42.01	47.16	36.31	41.83
InternVL3-8B 2026.01		40.05	41.15	32.77	37.99
InternVL3-14B 2026.01		39.83	40.84	36.35	39.01
Qwen2.5-VL-32B-Instruct 2026.01		38.41	41.3	34.29	38
InternVL3-2B 2026.01		37.71	35.99	26.19	33.3
Qwen2.5-VL-7B-Instruct 2026.01		36.26	37.54	32.84	35.55
Qwen2.5-VL-3B-Instruct 2026.01		35.46	30.81	25.29	30.52
InternVL3-38B 2026.01		34.27	38.48	40.65	37.8
InternVL3-1B 2026.01		33.26	20.96	22.36	25.53
Llama-3.2-11B-Vision-Instruct 2026.01		31.59	32.91	29.34	31.28
Llama-3.2-90B-Vision-Instruct 2026.01		27.26	26.33	27.72	27.1
Qwen2.5-VL-72B-Instruct 2026.01		23.78	27.67	50.76	34.07