SOTA Vision-Language Understanding benchmarks and papers with code

Benchmarks

Dataset Name	SOTA Method	Metric
MMBench	Qwen3-VL-4B-Instruct	Accuracy88.7	88	18d ago
MM-Vet	REVIS	Total Score72.16	43	29d ago
Vision-Language Benchmark Suite Aggregate		Aggregate Performance (%)100	34	3mo ago
MVista	ZPPO	Accuracy80.5	28	1mo ago
MMBench cn		Accuracy60.6	27	1mo ago
MME	Vanilla	Average Score100	18	1mo ago
Naturalbench	SemDeDup	General Score13.2	13	2mo ago
Winoground (test)	LLaVA-OV	Text Score76.75	12	1mo ago
VStarBench	Qwen3-VL-8B	Accuracy83.77	12	2mo ago
MVerse VI	Qwen3-VL-8B	Accuracy43.78	12	2mo ago
MVerse VO	Qwen3-VL-8B	Accuracy39.97	12	2mo ago
NegBench VOC2007	InternVL-3.0-8B	Accuracy95.37	11	2mo ago
NegBench COCO	InternVL-3.0-8B	Accuracy93.19	11	2mo ago
SEED-Bench Image	Vanilla	Average Accuracy100	11	3mo ago
Vision-Language Evaluation Suite MMB, MMStar, MMMU, Hallusion, AI2D, OCR, SEED, SQA (test val)	Qwen2-VL-7B (Teacher)	MMB Score80.7	10	4mo ago
MMStar (test)	ETTC	Accuracy63.73	7	1mo ago
MMBench CN v1.1	Qwen3-Omni	Accuracy87.7	5	2mo ago
MMBench EN v1.1	MiniCPM-o 4.5	Accuracy89	5	2mo ago
Winoground	CROCScore	Text Accuracy61.5	5	3mo ago
MMBench (test)		Overall Accuracy78.8	4	1mo ago
Quantiphy	Phi-4-Multimodal	Average MRA27.4	4	2mo ago
MMBench (dev)	Prompt Highlighter	Accuracy69.7	4	4mo ago
MME Perception	Prompt Highlighter	MME Score1,552.5	4	4mo ago
MMStar	SmoothSMoE annealed (k=2)	Accuracy42.64	3	1mo ago
Vision-Language Evaluation Suite (ChartQA, DocVQA, AI2D, VQA, AndroidControl, CountBenchQA)	Our Method	ChartQA Accuracy68.1	2	3mo ago

Showing 25 of 27 rows