How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings
Document parsing models are improving rapidly, but existing benchmarks are still insufficient for a comprehensive evaluation of their capabilities. We introduce PureDocBench, a benchmark with 10 domains, 66 subcategories, and 1,475 pages. Each page has matched Clean, Digital, and Real versions, covering clean renders, simulated degradation, and real capture or sharing conditions. The resulting 4,425 images have source-derived annotations for text, formulas, tables, and reading order, checked through automated validation and human review. Keeping page content and annotations fixed allows direct comparisons across image conditions. We evaluate 58 models, including pipeline-based parsers, en...
| Authors | Zhiheng Li, Zongyang Ma, Jiaxian Chen et al. |
|---|---|
| Primary category | cs.CV |
| Published | 8 May 2026 |
| Last updated at source | 27 September 2026 |
Connected in the knowledge graph
- zhihengli-casia/puredocbenchDataset · cited by
Models and datasets referencing this paper
- zhihengli-casia/puredocbenchDataset
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.