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

Video Question Answering research benchmarks – Page 11 · SOTA2 Research

Video Question Answering

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

STAR v1.0 (test)

73.7Interaction Accuracy

10

Feb 26, 2026

SEED-Bench Video

53.2Accuracy

10

Feb 26, 2026

95.4Accuracy

10

Feb 26, 2026

95.5Accuracy

10

Feb 26, 2026

42.8Accuracy

10

Feb 26, 2026

55.3ME

9

Jul 7, 2026

43.7ME Score

9

Jul 7, 2026

Minerva-Ego 1.0 (test)

Human performance

91.8MCQ Accuracy

9

May 18, 2026

Video-MME no subtitles v2 (test)

14.91Non-Linear Accuracy

9

May 12, 2026

Baseline (Qwen3-VL-8B-Instruct)

49.8VYR

9

Apr 21, 2026

63.2VYR

9

Apr 21, 2026

HD-EPIC 1 (test)

Sample Human Baseline

96.7Recipe

9

Apr 14, 2026

83.6Delta Knowledge

9

Mar 13, 2026

Finding Dory (dev)

27.5Overall Relaxed Accuracy

9

Mar 5, 2026

MLVU 2024 (test)

54.9Accuracy

9

Feb 26, 2026

75.4Accuracy

9

Feb 26, 2026

75.4Accuracy

9

Feb 26, 2026

NExT-QA v1 (val)

73.8Accuracy

9

Feb 26, 2026

VideoChatGPT (test)

2.96CI

9

Feb 26, 2026

Baseline+CoT+SFT

30.9VYR

9

Apr 21, 2026

56.6Accuracy

9

Feb 26, 2026

86.6Accuracy

9

Feb 26, 2026

VCGBench ActivityNet

3.27CI

9

Feb 26, 2026

EgoTaskQA indirect

47.82Accuracy (World)

9

Feb 26, 2026

48.6Accuracy

9

Feb 26, 2026

Page 11 of 18

Popular datasets

Video-MME VideoMME NExT-QA EgoSchema TGIF-QA

Follow for updates

Join SOTA2 Contact