InfoChartQA is a benchmark for evaluating multimodal large language models (MLLMs) on infographic charts enriched with pictorial visual elements like pictograms and icons. It features 5,948 pairs of infographic and plain charts that share the same underlying data but differ in visual style, enabling controlled comparisons. The dataset contains a total of 58,857 questions, including 50,920 text-based and 7,937 visual-element-based questions designed to probe model understanding of both content and complex visual design. Our analysis of 20 MLLMs reveals significant performance drops on infographic charts, highlighting key challenges and new research directions.
You can find our dataset on huggingface: InfoChartQA Dataset
Each question entry is arranged as follows. Note that for visual questions, there may be some extra input figures, which are cropped from the orginal figure. We present their bboxes in "extra_input_figure_bboxes".
{
"question_id": id of the question,
"question_type_name": question type name, for example: "extreme" questions,
"question_type_id": question type id, this is only used for evaluation! For example: 72 means "extreme" questions,
"figure_id": id of the figure,
"question": question text,
"answer": ground truth answer,
"instructions": instructions,
"url": url of the input image,
"extra_input_figure_ids": ids of the extra input figures,
"extra_input_figure_bboxes": bbox of the extra input figures, in [x,y,w,h] format w/o normalization.
"difficulty": difficulty level,
"chart_type": chart_type,
}
Each question is built by:
input_image: item["url"] (may need to download for models that don't support url input)
extra_input_image: Cropped input_image using item["extra_input_figure_bboxes"],
input_text: item["question"] + item["instructions"] (if any)
where item is an entry of the dataset.
For detailed evaluation instructions and usage, please refer to the Evaluation.
- InfoChartQA: A Benchmark for Multimodal Question Answering on Infographic Charts
Tianchi Xie, Minzhi Lin, Mengchen Liu, Yilin Ye, Changjian Chen, Shixia Liu
If you use our work and are inspired by our work, please cite:
@misc{lin2025infochartqa,
title={InfoChartQA: A Benchmark for Multimodal Question Answering on Infographic Charts},
author={Tianchi Xie and Minzhi Lin and Mengchen Liu and Yilin Ye and Changjian Chen and Shixia Liu},
year={2025},
eprint={2505.19028},
url={https://arxiv.org/abs/2505.19028},
}
Our original data contributions (all data except the charts) are distributed under the CC BY-SA 4.0 license. The copyright of the charts belong to the original authors.
-
OrionBench: A Benchmark for Chart and Human-Recognizable Object Detection in Infographics
Paper | Code | Dataset -
ChartGalaxy: A Dataset for Infographic Chart Understanding and Generation
paper | Code | Dataset
If you have any questions about this work, please contact us using the following email address: [email protected].