T2I-ReasonBench: Benchmarking Reasoning-Informed Text-to-Image Generation

Kaiyue Sun, Rongyao Fang, Chengqi Duan, Xian Liu, Xihui Liu
8/24/2025
cs.CV

No official code found. Best implementation: KaiyueSun98/T2I-ReasonBench (37 stars, last commit 2025-09-16).

Abstract

We propose T2I-ReasonBench, a benchmark evaluating reasoning capabilities of text-to-image (T2I) models. It consists of four dimensions: Idiom Interpretation, Textual Image Design, Entity-Reasoning and Scientific-Reasoning. We propose a two-stage evaluation protocol to assess the reasoning accuracy and image quality. We benchmark various T2I generation models, and provide comprehensive analysis on their performances.

View on arXivView PDF

Code Implementations(1)

KaiyueSun98/T2I-ReasonBenchCommunity70%

T2I-ReasonBench: Benchmarking Reasoning-Informed Text-to-Image Generation

372Shell, PythonAug 18, 20251 years ago

Similar Papers

GenExam: A Multidisciplinary Text-to-Image Exam

88% match1 repo692025

UniGenBench++: A Unified Semantic Evaluation Benchmark for Text-to-Image Generation

88% match1 repo1392025

T2I-FactualBench: Benchmarking the Factuality of Text-to-Image Models with Knowledge-Intensive Concepts

87% match1 repo82024

Easier Painting Than Thinking: Can Text-to-Image Models Set the Stage, but Not Direct the Play?

86% match1 repo522025

Reflect-DiT: Inference-Time Scaling for Text-to-Image Diffusion Transformers via In-Context Reflection

86% match1 repo562025

ChatGen: Automatic Text-to-Image Generation From FreeStyle Chatting

86% match1 repo332026

Cite this paper

@article{sun2025tireasonbench,
  title  = {T2I-ReasonBench: Benchmarking Reasoning-Informed Text-to-Image Generation},
  author = {Kaiyue Sun and Rongyao Fang and Chengqi Duan and Xian Liu and Xihui Liu},
  year   = {2025},
  eprint = {2508.17472},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2508.17472v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2508.17472.svg)](https://codeofpaper.com/paper/2508.17472)

Discussion