---
title: "Published Document Processing Benchmarks | Extend"
description: "Published document processing benchmarks for parsing, structured extraction, and splitting, with public datasets, methodology, and reproducible source material."
canonical: https://www.extend.ai/benchmarks
---

# Published Document Processing Benchmarks | Extend

benchmarks

## Benchmarks for the hardest parts of document processing

Published document processing benchmarks for parsing, structured extraction, and splitting. Each report states its corpus, metric, comparison scope, methodology, and public data or code.

## Benchmarks

benchmark blogs. Public benchmarks for document parsing, extraction, and splitting, with published methodology and source data. Each card separates key results, benchmark details, and sources.

### RealDoc-Bench

Parsing · [RealDoc-Bench](https://www.extend.ai/resources/realdocbench)

Measures whether parsers deliver accurate layouts, preserve reading order, and enable agents to correctly answer objective questions against real-world documents.

**RealDoc-Bench key results**

| Result | Value | Detail |
| --- | --- | --- |
| Q&A accuracy | 95.7% | Extend Parse 2.0 |
| Layout F1 | 0.847 | Adjusted F1 |

**Benchmark details for RealDoc-Bench**

| Field | Detail |
| --- | --- |
| Test corpus | 1,500 layout samples; 1,359 document Q&A prompts across 581 documents in four regulated industries |
| Metric | Adjusted F1 for layout; field-level document Q&A accuracy |
| Comparison | Same-run comparison with LlamaParse, Reducto, Azure Document Intelligence, and AWS Textract |
| Published | May 26, 2026 |

**Sources**

- [Research paper — arXiv](https://arxiv.org/abs/2606.07401)
- [Source code — GitHub repository](https://github.com/extend-hq/realdoc-bench)
- [Layout dataset — Hugging Face](https://huggingface.co/datasets/Extend-AI/RealDoc-Bench-Layout/)
- [Document Q&A dataset — Hugging Face](https://huggingface.co/datasets/Extend-AI/RealDoc-Bench)

[Read RealDoc-Bench](https://www.extend.ai/resources/realdocbench)

### LongArray-Extract

Extraction · [LongArray-Extract](https://www.extend.ai/resources/long-array-extraction-benchmark)

Tests whether extraction systems preserve cardinality and return complete, schema-faithful arrays when the output grows from a dozen of rows to thousands.

**LongArray-Extract key results**

| Result | Value | Detail |
| --- | --- | --- |
| Mean accuracy | 99.2% | Extend across 45 PDFs |
| Run completion | 100% | 45 of 45 PDFs completed |

**Benchmark details for LongArray-Extract**

| Field | Detail |
| --- | --- |
| Test corpus | 45 financial, clinical, and legal PDFs with repeated arrays of 27 to about 2,200 records |
| Metric | Mean per-document extraction accuracy, with failed and timed-out runs scored as zero |
| Comparison | Document-AI platforms: Reducto, Pulse, LlamaParse; Raw-model providers: Anthropic, Google, OpenAI |
| Published | June 2, 2026 |

**Sources**

- [Extraction dataset — Hugging Face](https://huggingface.co/datasets/Extend-AI/LongArray-Extract)

[Read LongArray-Extract](https://www.extend.ai/resources/long-array-extraction-benchmark)

### PoliTax Split

Splitting · [PoliTax Split](https://www.extend.ai/resources/document-splitting-benchmark)

Evaluates document splitting on long, compound tax filings where frontier models miss subtle boundaries across hundreds of pages.

**PoliTax Split key results**

| Result | Value | Detail |
| --- | --- | --- |
| Best harness F1 | 72.48% | Claude Opus 4.6 |
| Recall lift | 17-44 | Points across models |

**Benchmark details for PoliTax Split**

| Field | Detail |
| --- | --- |
| Test corpus | The 30 largest compound documents in the public PoliTax corpus |
| Metric | Boundary-detection F1 |
| Comparison | Same-model comparison between the Extend splitting harness and direct frontier-model use |
| Published | March 30, 2026 |

**Sources**

- [Splitting dataset — Hugging Face](https://huggingface.co/datasets/Extend-AI/PoliTax-Split)

[Read PoliTax Split](https://www.extend.ai/resources/document-splitting-benchmark)