# StructEval: Benchmarking LLMs' Capabilities to Generate Structural Outputs

> Evaluates generation and conversion of structured outputs, including HTML, React and SVG, with separate checks for format adherence and structural correctness.

- Authors: [Authors: see the paper](https://arxiv.org/abs/2505.20139)
- Venue: arXiv
- Status: Preprint
- Publication / posting date: May 26, 2025
- Date source: Original arXiv first-submission date; reviewed preprint edition retained.
- First arXiv submission: May 26, 2025
- arXiv: 2505.20139
- Reviewed arXiv version: v3
- Sources reviewed: 2026-10-20
- Added to the catalog: 2026-11-01
- Original paper: <https://arxiv.org/abs/2505.20139>
- PDF: <https://arxiv.org/pdf/2505.20139v3>
- Code: <https://github.com/TIGER-AI-Lab/StructEval>
- Project: <https://tiger-ai-lab.github.io/StructEval/>
- BibTeX: <https://awesomerea.org/papers/citations/paper-structeval.bib>
- Catalog page: <https://awesomerea.org/papers/paper-structeval/>
- Record JSON: <https://awesomerea.org/papers/paper-structeval/record.json>
- Chinese version: [简体中文](https://awesomerea.org/zh/papers/paper-structeval/index.md)

## Scope and limitations

Not a software reverse-engineering or GUI-navigation benchmark. Structured-output scores do not prove recovered source identity, backend reconstruction or behavioral equivalence; aggregate scores are not fully correct-output percentages.

## Why it fits this collection

A supporting method for checking structured and rendered artifacts produced during frontend reconstruction; parseability alone does not establish correct content.

## Publication and review notes

This record describes the original arXiv edition. Supporting method.

## Primary sources

1. <https://arxiv.org/abs/2505.20139v3>
2. <https://github.com/TIGER-AI-Lab/StructEval>
3. <https://tiger-ai-lab.github.io/StructEval/>

## About this entry

Reviewed against the paper and its primary sources. Cite the original authors using the linked paper or BibTeX.
Original editorial content: MIT. Linked projects, papers and assets retain their own rights and licenses.