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 21 · SOTA2 Research

Text-to-Image Generation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Text-to-Image Evaluation Prompts

SenseFlow-SD3.5 Large

0.333CLIP Score

8

Mar 4, 2026

30.37PSNR

8

Feb 26, 2026

GenEval zero-shot

Meissonic + Co-GRPO

0.55GenEval Score

8

Feb 26, 2026

0.91P Score

8

Feb 26, 2026

COCO 10k-sample 2017

90Precision (P)

8

Feb 26, 2026

Stable Diffusion Alignment Prompts 1.5 (test)

Adjoint Matching

0.787ImageReward

8

Feb 26, 2026

GenEval 11 (test)

0.275Overall Score

8

Feb 26, 2026

MJHQ-30K 18 (test)

16.2FID

8

Feb 26, 2026

84Accuracy

8

Feb 26, 2026

2.607NIQE

8

Feb 26, 2026

DSG prompts (test)

78DSGmPLUG Score

8

Feb 26, 2026

Physically Implausible Scenarios

78DSGmPLUG

8

Feb 26, 2026

GenEval original (test)

99Single Object Accuracy

8

Feb 26, 2026

Prompt Reinjection

1.319ImageReward

8

Feb 26, 2026

Four Positive Affordances Evaluation Set

3.86Faithfulness

8

Feb 26, 2026

Concept-Affordance Distant

4.59Novelty

8

Feb 26, 2026

Concept-Affordance Close

4.46Novelty

8

Feb 26, 2026

LAION 10K 5B (test)

PixArt-Σ w/ DiffCR-LT

10.68FID

8

Feb 26, 2026

0.926Similarity (95th Percentile)

8

Feb 26, 2026

COCO 5k SDXL 1024px 2017 (val)

33.16CLIP Score

8

Feb 26, 2026

ImagenHub (test)

0.29CLIP-T Score

8

Feb 26, 2026

DrawBench (test)

0.427CLIP-Score

8

Jul 2, 2026

curated 30K High-Quality Evaluation Dataset

7.38FID

8

Feb 26, 2026

Multi30K Task2 DE (test)

ENSAD + Fine-Tune G

14.93FID

8

Feb 26, 2026

93.13Spatial Fidelity (Human)

8

Feb 26, 2026

Page 21 of 39

Popular datasets

GenEval MS-COCO DrawBench COCO T2I-CompBench

Follow for updates

Join SOTA2 Contact