Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

General Multimodal Understanding research benchmarks · SOTA2 Research

General Multimodal Understanding

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Qwen3-VL-32B-Instruct

56.9Accuracy

31

Jun 9, 2026

73.5Accuracy

29

Jun 30, 2026

88.1Accuracy

17

Jun 30, 2026

73.6Accuracy

17

Jun 30, 2026

General Multimodal Evaluation Suite (MMMU, MMBench, MME, ChartQA, AI2D, HallBench)

72.6MMMU (Val)

14

Feb 26, 2026

Aggregated Benchmarks

71Average Score

13

May 14, 2026

Combined 9 Benchmarks

LLaVA-NeXT Vanilla 13B

100Average Accuracy

13

Mar 4, 2026

66.63Normalized Score

12

May 8, 2026

General Benchmarks

74Average Score

12

Feb 26, 2026

Qwen3-VL-30B-CoT

77.6Accuracy

10

Mar 25, 2026

Vision-Language Benchmark Suite

SmoothSMoE (k=2.5)

55.07Average Score

3

Jun 18, 2026

Popular datasets

MMMU MME ChartQA BLINK MM

Follow for updates

Join SOTA2 Contact