DocParsingBench: leaderboard
Systems turn 1,400 pages of real business documents (finance, legal, research, manufacturing, education; Chinese and English) into Markdown; the README does not state how many pages each system was scored on. The score is DPB, SoMark's overall mark built from text edit distance, formula CDM and table TEDS with weights 0.5, 0.3 and 0.2. Unlike OmniDocBench, the pages come from production documents with stamps, scans and handwriting and the order of segments is matched rather than fixed.
Metric: DPB overall score (0-1). Source: github.com. Saturation forecast: Rough model projection: around 2026. 17 models tracked. Updated 16 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | OXR-1.0 | 0.8776 | |
| 2 | PaddleOCR-VL-1.6 | 0.8517 | |
| 3 | Jina-OCR-v1 | 0.8446 | |
| 4 | FireRed-OCR | 0.8405 | |
| 5 | TeleOCR | 0.8313 | |
| 6 | Dots.mocr | 0.8306 | |
| 7 | MinerU2.5-Pro | 0.8278 | |
| 8 | Infinity-Parser2-Pro | 0.8263 | |
| 9 | OvisOCR2 | 0.8226 | |
| 10 | Chandra-OCR-2 | 0.822 | |
| 11 | Qianfan-OCR | 0.8145 | |
| 12 | ABot-OCR | 0.8042 | |
| 13 | GLM-OCR | 0.8034 | |
| 14 | Youtu-Parsing | 0.7625 | |
| 15 | HunyuanOCR 1.5 | 0.7535 |
Interactive version: theaggregate.ai/benchmark?slug=docparsingbench · How It Works · Data refreshed daily, snapshot 2026-11-16.