ChartMuseum: Testing Visual Reasoning Capabilities of Large Vision-Language Models

Liyan Tang, Grace Kim, Xinyu Zhao, Thom Lake, Wenxuan Ding, Fangcong Yin, Prasann Singhal, Manya Wadhwa, Zeyu Leo Liu, Zayne Rea Sprague, Ramya Namuduri, Bodun Hu, Juan Diego Rodriguez, Puyuan Peng, Greg Durrett

Keywords: chart understanding, LVLM, reasoning

Abstract: Chart understanding presents a unique challenge for large vision-language models (LVLMs), as it requires the integration of sophisticated textual and visual reasoning capabilities. However, current LVLMs exhibit a notable imbalance between these skills, falling short on visual reasoning that is difficult to perform in text. We conduct a case study using a synthetic dataset solvable only through visual reasoning and show that model performance degrades significantly with increasing visual complexity, while human performance remains robust. We then introduce *ChartMuseum*, a new Chart Question Answering (QA) benchmark containing 1,162 expert-annotated questions spanning multiple reasoning types, curated from real-world charts across 184 sources, specifically built to evaluate complex visual and textual reasoning. Unlike prior chart understanding benchmarks---where frontier models perform similarly and near saturation---our benchmark exposes a substantial gap between model and human performance, while effectively differentiating model capabilities: although humans achieve 93% accuracy, the best-performing model Gemini-2.5-Pro attains only 63.0%, and the leading open-source LVLM Qwen2.5-VL-72B-Instruct achieves only 38.5%. Moreover, on questions requiring primarily visual reasoning, *all* models experience a 35%-55% performance drop from text-reasoning-heavy question performance. Lastly, our qualitative error analysis reveals specific categories of visual reasoning that are challenging for current LVLMs. Both ChartMuseum and the evaluation code are available at [https://github.com/Liyan06/ChartMuseum](https://github.com/Liyan06/ChartMuseum).

Dataset URL: https://huggingface.co/datasets/lytang/ChartMuseum

Code URL: https://github.com/Liyan06/ChartMuseum

Primary Area: Datasets & Benchmarks for applications in language modeling and vision language modeling

Submission Number: 2133