Perception on MMStar latest (test) benchmark leaderboard · SOTA2 Research

Perception on MMStar latest (test)

67.2CPGPT-4o

Evaluation Results

Method	Links
GPT-4o2024.06		67.2	51.6	63.2	56.4	52	36.8	54.5
GPT-4v2024.06		67.2	43.6	56.8	46.8	42	30.8	47.9
InternVL-Chat-v1.52024.06		65.6	42.4	56.4	50.8	46.8	28	48.3
LLaVA-NeXT (Yi-34B)2024.06		64.4	37.2	50	34	39.2	24.8	41.6
GPT-4o2024.06		64	41.6	54.4	51.6	47.6	31.2	48.4
LLaVA-NeXT (Vicuna-13B)2024.06		63.2	38.8	49.2	38.8	37.6	23.2	41.8
LLaVA-NeXT (Vicuna-7B)2024.06		63.2	28.8	46.4	40.8	34	22.8	39.3
LLaVA-NeXT (Mistral-7B)2024.06		62.8	36.4	53.6	42	35.2	26.4	42.7
GPT-4v2024.06		62.4	33.2	51.2	39.2	44.4	32.8	43.9
Qwen-VL-Max2024.06		62.4	28.8	49.2	45.2	37.2	21.2	40.7
Qwen-VL-Max2024.06		62.4	36	54.8	46.4	38	24	43.6
GeminiPro-V2024.06		61.2	35.2	53.2	47.6	48.4	19.2	44.1
DeepSeek-VL-7B2024.06		61.2	34.8	54.4	42.8	45.2	25.2	43.9
MiniCPM-V-22024.06		61.2	30.4	52	44.4	40.8	22.8	41.9
InternLM-XComposer22024.06		60.8	39.2	58.4	47.6	42.4	19.6	44.7
LLaVA-NeXT (Mistral-7B)2024.06		60.8	30.8	49.6	37.6	36.8	20.4	39.3
LLaVA-NeXT (Vicuna-13B)2024.06		60.8	38	48.8	35.2	43.6	20.8	41.2
LLaVA-NeXT (Yi-34B)2024.06		60	41.6	52.4	44	40.8	25.2	44
InternVL-Chat-v1.52024.06		59.2	33.2	51.2	42.8	40.4	30.4	42.9
GeminiPro-V2024.06		57.2	36	51.6	43.2	45.6	23.2	42.8
DeepSeek-VL-7B2024.06		57.2	31.6	49.6	42	43.6	24	41.3
LLaVA-NeXT (Vicuna-7B)2024.06		56.8	37.6	47.2	42.4	36.4	21.6	40.3
InternLM-XComposer22024.06		56	32	54.8	33.6	37.2	20.8	39.1
MiniCPM-V-22024.06		55.6	28.4	49.2	39.6	38.4	19.2	38.4
LLaVA-v1.5-13B2024.06		54.8	28.8	48.8	37.6	32	20	37
mPLUG-Owl22024.06		53.2	34.4	45.2	38.8	38.8	23.6	39
LLaVA-v1.5-7B2024.06		50.8	31.6	50.4	38.8	32	25.2	38.1
LLaVA-v1.5-7B2024.06		48.8	29.6	48	34.4	27.6	17.2	34.3
mPLUG-Owl22024.06		47.2	24	45.2	33.6	28.8	18.8	32.9
LLaVA-v1.5-13B2024.06		45.2	28.4	45.2	35.6	30	17.2	33.6