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

Visual Question Answering research benchmarks – Page 40 · SOTA2 Research

Visual Question Answering

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

89Perception Score

4

Mar 17, 2026

67.2Lingo-Judge

4

Mar 17, 2026

InfoGraphic-VQA (val)

Qwen2.5-VL-7B + Cont. Reward

81.06Accuracy

4

Mar 16, 2026

COCO 300 sampled images (test)

100Clean Accuracy

4

Mar 16, 2026

75.4Score

4

Mar 11, 2026

This-is-my Single Concept (test)

92Accuracy

4

Mar 11, 2026

Yo’LLaVA Single Concept (test)

97.6Accuracy

4

Mar 11, 2026

VizWiz-QA (test)

70.1Accuracy

4

Feb 27, 2026

78.5Accuracy

4

Feb 27, 2026

61.9GQA Score

4

Feb 27, 2026

78.7Accuracy

4

Feb 27, 2026

79.99RAD (Y/N ← C)

4

Feb 26, 2026

80.2Accuracy

4

Feb 26, 2026

76.6Accuracy (Aut)

4

Feb 26, 2026

69.5Accuracy

4

Feb 26, 2026

65.3Accuracy (MCQ)

4

Feb 26, 2026

CTIS-Bench (test)

0.523H.F.

4

Feb 26, 2026

64.9Accuracy

4

Feb 26, 2026

62.9Score

4

Feb 26, 2026

77.8Accuracy

4

Feb 26, 2026

Video-MMMU 1.0 (sampled examples)

86Acc

4

Feb 26, 2026

WorldSense sampled examples 1.0

60Accuracy

4

Feb 26, 2026

VQA 100 prompts

52.54Accuracy

4

Feb 26, 2026

OCRBench (test)

Proprietary API SOTA (SenseTime, 2024)

89.4Score

4

Feb 26, 2026

MathVision full (test)

Proprietary API SOTA (Google, 2024)

43.6Accuracy

4

Feb 26, 2026

Page 40 of 46

Popular datasets

VQA GQA TextVQA Clevr Vizwiz

Follow for updates

Join SOTA2 Contact