WildTableBench: Benchmarking Multimodal Foundation Models on Table Understanding In the Wild

Hugging Face Daily Papers Papers

Summary

WildTableBench introduces the first question-answering benchmark for real-world table images, revealing that existing multimodal foundation models struggle significantly with structural perception and numerical reasoning, with only one model exceeding 50% accuracy.

Using multimodal foundation models to analyze table images is a high-value yet challenging application in consumer and enterprise scenarios. Despite its importance, current evaluations rely largely on structured-text tables or clean rendered images, leaving the visual complexity of in-the-wild table images underexplored. Such images feature varied layouts and diverse domains that demand sophisticated structural perception and numerical reasoning. To bridge this gap, we introduce WildTableBench, the first question-answering benchmark for naturally occurring table images from real-world settings. WildTableBench comprises 402 high-information-density table images collected from online forums and websites across diverse domains, together with 928 manually annotated and verified questions spanning 17 subtypes across five categories. We evaluate 21 frontier proprietary and open-source multimodal foundation models on this benchmark. Only one model exceeds 50% accuracy, while all remaining models range from 4.1% to 49.9%. We further conduct diagnostic analyses to characterize model failures and reveal persistent weaknesses in structural perception and reasoning. These results and analyses provide useful insights into current model capabilities and establish WildTableBench as a valuable diagnostic benchmark for table image understanding.
Original Article
View Cached Full Text

Cached at: 05/15/26, 08:27 PM

Paper page - WildTableBench: Benchmarking Multimodal Foundation Models on Table Understanding In the Wild

Source: https://huggingface.co/papers/2605.01018 Published on May 1

·

Submitted byhttps://huggingface.co/jzhuang

HJZon May 15

Abstract

WildTableBench is introduced as the first question-answering benchmark for real-world table images, revealing significant challenges in structural perception and numerical reasoning for existing multimodal models.

Usingmultimodal foundation modelsto analyzetable imagesis a high-value yet challenging application in consumer and enterprise scenarios. Despite its importance, current evaluations rely largely on structured-text tables or clean rendered images, leaving the visual complexity of in-the-wildtable imagesunderexplored. Such images feature varied layouts and diverse domains that demand sophisticatedstructural perceptionandnumerical reasoning. To bridge this gap, we introduce WildTableBench, the firstquestion-answering benchmarkfor naturally occurringtable imagesfrom real-world settings. WildTableBench comprises 402 high-information-densitytable imagescollected from online forums and websites across diverse domains, together with 928 manually annotated and verified questions spanning 17 subtypes across five categories. We evaluate 21 frontier proprietary and open-sourcemultimodal foundation modelson this benchmark. Only one model exceeds 50% accuracy, while all remaining models range from 4.1% to 49.9%. We further conduct diagnostic analyses to characterize model failures and reveal persistent weaknesses instructural perceptionand reasoning. These results and analyses provide useful insights into current model capabilities and establish WildTableBench as a valuable diagnostic benchmark for table image understanding.

View arXiv pageView PDFProject pageGitHub0Add to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2605.01018 in a model README.md to link it from this page.

Datasets citing this paper1

#### jzhuang/WildTableBench Viewer• Updated10 days ago • 928 • 56

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2605.01018 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

MulTaBench: Benchmarking Multimodal Tabular Learning with Text and Image

Hugging Face Daily Papers

Introduces MulTaBench, a benchmark of 40 datasets for multimodal tabular learning with text and image modalities, demonstrating that task-specific embedding tuning improves performance over frozen pretrained embeddings, particularly when modalities provide complementary predictive signals.

TabularMath: Understanding Math Reasoning over Tables with Large Language Models

arXiv cs.CL

TabularMath introduces a benchmark and AutoT2T framework for evaluating LLMs' mathematical reasoning over tabular data, revealing that table complexity, data quality, and modality significantly impact model performance. The study addresses a gap in LLM evaluation by systematically assessing robustness to incomplete or inconsistent table information in real-world scenarios.