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

Text-to-Image Generation research benchmarks – Page 24 · SOTA2 Research

Text-to-Image Generation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

0.309CLIP Score

7

Apr 7, 2026

80.9ASR

7

Feb 26, 2026

Semantic Erasure and Horizontal Injection

17.597PickScore

7

Feb 26, 2026

Out-of-Domain Evaluation Set

FLUX.1-dev + UnifiedReward-Flex

36.25CLIP Score

7

Feb 26, 2026

GenEval out-of-domain

FLUX.1-dev + UnifiedReward-Flex

69.62Semantic Consistency

7

Feb 26, 2026

T2I-CompBench out-of-domain

FLUX.1-dev + UnifiedReward-Flex

51.37Semantic Consistency

7

Feb 26, 2026

UniGenBench++ in-domain

FLUX.1-dev + UnifiedReward-Flex

73.95Semantic Consistency

7

Feb 26, 2026

3.06UnifiedReward

7

Feb 26, 2026

LAION-10k Scenario 1 (test)

0.65Similarity (95pc)

7

Feb 26, 2026

24.93FID (1K)

7

Feb 26, 2026

In-the-wild image color condition

50.9FID

7

Feb 26, 2026

Sampled color condition (Manual)

Stable Diffusion

72.6FID

7

Feb 26, 2026

mmDiT-XL/2 (ours)

6.27FID

7

Feb 26, 2026

Custom Compositional Prompts 1.0 (test)

SynGen + TEBOpt

56.52 Objects Exist Score

7

Feb 26, 2026

T2I-Compbench Count

48Human Accuracy

7

Feb 26, 2026

50YOLOv9 Accuracy

7

Feb 26, 2026

LAION-5B FI (test)

ENSAD + Fine-Tune G

16.23FID

7

Feb 26, 2026

MT Bench 90 prompts (test)

808Total Wins

7

Feb 26, 2026

DiffusionDB Real User Prompts 466 prompts (test)

1,508Win Count

7

Feb 26, 2026

COCO caption (val)

Versatile Diffusion

11.1FID

7

Feb 26, 2026

Muse (base + super-res)

6.06FID

7

Feb 26, 2026

Beautiful-Prompt (test)

20.86PickScore

7

Feb 26, 2026

Race attributes Stable Diffusion v2.1

0.15Bias-O

6

Jul 2, 2026

Gender attributes Stable Diffusion v2.1

0.26Bias-O

6

Jul 2, 2026

GenEval short prompts original

0.48GenEval Score

6

Jul 1, 2026

Page 24 of 39

Popular datasets

GenEval MS-COCO DrawBench COCO T2I-CompBench

Follow for updates

Join SOTA2 Contact