Large Language Models (LLMs) are increasingly being adopted as tools for learning; however, most tools remain text-only, limiting their usefulness for domains where visualizations are essential, such as mathematics. Recent work shows that LLMs are capable of generating code that compiles to educational figures, but a major bottleneck remains: scalable evaluation of these diagrams. We address this by proposing DiagramIR: an automatic and scalable evaluation pipeline for geometric figures. Our method relies on intermediate representations (IRs) of LaTeX TikZ code. We compare our pipeline to other evaluation baselines such as LLM-as-a-Judge, showing that our approach has higher agreement with human raters. This evaluation approach also enables smaller models like GPT-4.1-Mini to perform comparably to larger models such as GPT-5 at a 10x lower inference cost, which is important for deploying accessible and scalable education technologies.
@article{arxiv.2511.08283,
title = {DiagramIR: An Automatic Pipeline for Educational Math Diagram Evaluation},
author = {Vishal Kumar and Shubhra Mishra and Rebecca Hao and Rizwaan Malik and David Broman and Dorottya Demszky},
journal= {arXiv preprint arXiv:2511.08283},
year = {2025}
}