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 12 · SOTA2 Research

Text-to-Image Generation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

COCO 5,000 images (val)

2.2FID

14

May 8, 2026

0.838TA Score

14

Mar 25, 2026

UniGenBench++ Long

76.22Score

14

Mar 13, 2026

UniGenBench++ Short

69.56Score

14

Mar 13, 2026

TIIF Long Prompts

89.7Basic Following (Avg)

14

Apr 15, 2026

Flickr 30K 5,000 samples

Original (SDXL)

33.95FID

14

Mar 6, 2026

MS COCO 5,000 samples

Original (SDXL)

27.43FID

14

Mar 6, 2026

80.96VQAScore

14

Mar 4, 2026

Momentum Guidance

31.47HPSv2.1 Score

14

Feb 26, 2026

Text-to-Image Preference Evaluation Suite (HPSv2.1, ImageReward, PickScore, Aes.Pred.v2.5, CLIP, Unified Reward) v2.1

0.361HPSv2.1

14

Feb 26, 2026

Dataset D2 (test)

Curriculum-DPO++

0.627Text Alignment

14

Feb 26, 2026

Dataset D1 (test)

0.77Text Alignment

14

Feb 26, 2026

T2I-CompBench 1.0 (test)

0.344CLIP Score

14

Feb 26, 2026

MS-COCO 256x256 (test)

6.61FID (30K)

14

Feb 26, 2026

MS-COCO 30K prompts (val)

3.22FID

14

Feb 26, 2026

6.36FID

14

Mar 10, 2026

9.78FID

14

May 4, 2026

GenExam Relaxed Score

87.8Math Score (GenExam Relaxed)

13

Jul 3, 2026

FLUX.1 dev native (test)

8.25Speedup

13

Jul 3, 2026

Qwen-Image 1664 x 928

6.95Latency (s)

13

Jul 1, 2026

T2I-CompBench standard

88.613D Spatial Score

13

Jun 30, 2026

MS-COCO Seen concepts

14.51FID

13

Jun 24, 2026

GenEval++ (test)

GPT Image + R3-Refiner

95Color Score

13

May 20, 2026

77.07Color Fidelity

13

Jul 7, 2026

FLUX.1 dev (test)

1.095ImageReward

13

May 19, 2026

Page 12 of 39

Popular datasets

GenEval MS-COCO DrawBench COCO T2I-CompBench

Follow for updates

Join SOTA2 Contact