# TexOCR: Advancing Document OCR Models for Compilable Page-to-LaTeX Reconstruction

arXiv: https://arxiv.org/abs/2604.22880 · HTML: https://mlsanity.com/paper/2604.22880/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 34 references that are in the map.
- First posted on arXiv 24 Apr 2026.

## Builds on (most cited first, 20 of 20)

- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone](https://mlsanity.com/paper/2404.14219/) (2024)
- [How Far Are We to GPT-4V? Closing the Gap to Commercial Multimodal Models with Open-Source Suites](https://mlsanity.com/paper/2404.16821/) (2024)
- [Nougat: Neural Optical Understanding for Academic Documents](https://mlsanity.com/paper/2308.13418/) (2023)
- [General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Model](https://mlsanity.com/paper/2409.01704/) (2024)
- [OmniDocBench: Benchmarking Diverse PDF Document Parsing with Comprehensive Annotations](https://mlsanity.com/paper/2412.07626/) (2024)
- [SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion](https://mlsanity.com/paper/2503.11576/) (2025)
- [olmOCR 2: Unit Test Rewards for Document OCR](https://mlsanity.com/paper/2510.19817/) (2025)
- [POINTS-Reader: Distillation-Free Adaptation of Vision-Language Models for Document Conversion](https://mlsanity.com/paper/2509.01215/) (2025)
- [Qianfan-OCR: A Unified End-to-End Model for Document Intelligence](https://mlsanity.com/paper/2603.13398/) (2026)
- [Infinity Parser: Layout Aware Reinforcement Learning for Scanned Document Parsing](https://mlsanity.com/paper/2510.15349/) (2025)
- [Calamari - A High-Performance Tensorflow-based Deep Learning Package for Optical Character Recognition](https://mlsanity.com/paper/1807.02004/) (2018)
- [READoc: A Unified Benchmark for Realistic Document Structured Extraction](https://mlsanity.com/paper/2409.05137/) (2024)
- [KITAB-Bench: A Comprehensive Multi-Domain Benchmark for Arabic OCR and Document Understanding](https://mlsanity.com/paper/2502.14949/) (2025)
- [Multimodal LLMs for OCR, OCR Post-Correction, and Named Entity Recognition in Historical Documents](https://mlsanity.com/paper/2504.00414/) (2025)
- [Table2LaTeX-RL: High-Fidelity LaTeX Code Generation from Table Images via Reinforced Multimodal Language Models](https://mlsanity.com/paper/2509.17589/) (2025)
- [LATTE: Improving Latex Recognition for Tables and Formulae with Iterative Refinement](https://mlsanity.com/paper/2409.14201/) (2024)
- [Context-Independent OCR with Multimodal LLMs: Effects of Image Resolution and Visual Complexity](https://mlsanity.com/paper/2503.23667/) (2025)
- [Complex Mathematical Expression Recognition: Benchmark, Large-Scale Dataset and Strong Baseline](https://mlsanity.com/paper/2512.13731/) (2025)

## Built on by (most cited first, 1 of 1)

- [TeXFix-Bench: An Empirically Grounded Multi-Format Benchmark for LLM-Based Document Source Repair](https://mlsanity.com/paper/2608.07617/) (2026)

## Abstract (arXiv)

Existing document OCR largely targets plain text or Markdown, discarding the structural and executable properties that make LaTeX essential for scientific publishing. We study page-level reconstruction of scientific PDFs into compilable LaTeX and introduce TexOCR-Bench, a benchmark, and TexOCR-Train, a large-scale training corpus, for this task. TexOCR-Bench features a multi-dimensional evaluation suite that jointly assesses transcription fidelity, structural faithfulness, and end-to-end compilability. Leveraging TexOCR-Train, we train a 2B-parameter model, TexOCR, using supervised fine-tuning (SFT) and reinforcement learning (RL) with verifiable rewards derived from LaTeX unit tests that directly enforce compilability and referential integrity. Experiments across 21 frontier models on TexOCR-Bench show that existing systems frequently violate key document invariants, including consistent section structure, correct float placement, and valid label-reference links, which undermines compilation reliability and downstream usability. Our analysis further reveals that RL with verifiable rewards yields consistent improvements over SFT alone, particularly on structural and compilation metrics.