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

Qwen-Image dataset and benchmarks · SOTA2 Research

Research Datasets

Qwen-Image

Benchmarks

Task NameDataset NameSOTA MethodSort

Task NameDataset NameSOTA ResultTrendResults

Text-to-Image Generation

Qwen-Image

1.295Image Reward

96

Single-task capability preservation

Qwen-Image

76.91DINO Score

24

Text-to-Image Generation

Qwen-Image 1664 x 928

6.95Latency (s)

13

Text-to-Image Generation

Qwen-Image Evaluation Set

22.4Latency (s)

12

Text-to-Image Generation

Qwen-Image native (test)

10.3Speedup

11

Image-level forgery detection

Qwen-Image

82.1Accuracy

9

Pixel-level image forgery localization

Qwen-Image (OOD)

48.2IoU

9

Identity Customization

Qwen-Image

0.703Similarity

8

Qwen-Image 12B 2.5 (test)

99.07Accuracy

7

Qwen-Image (test)

97.05Accuracy

7

Resolution extrapolation

Qwen-Image Direct extrapolation (test)

78.15FID

6

Style Customization

Qwen-Image

71Similarity

4

Text-to-Image Generation

Qwen-Image

28.13FID

3

Text-to-Image Generation

Qwen-Image-Lightning

10.91Latency (s)

3

Generative Diversity Evaluation

Qwen-Image

0.958DINOv3 Score

3

AI-generated image detection

Qwen-Image

94.1Accuracy

3

Human Evaluation

Qwen-Image rollout results

77.5Win Rate

2

Multimodal Inference

Qwen-Image (inference)

14.92Inference Latency (s)

2

Popular tasks

Text-to-Image Generation AIGI Detection Human Evaluation Image-level Forgery Detection Style Customization

Follow for updates

Join SOTA2 Contact