DocParsingBench - Formula: 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. This board is the display-formula part: CDM F1 of each matched formula, which compares rendered characters rather than LaTeX strings.

Metric: Display formula CDM F1 (0-1). Source: github.com. Saturation forecast: Rough model projection: around 2026. 17 models tracked. Updated 16 Nov 2026.

Top models

#ModelScoreOverall rank
1OXR-1.00.8576
2MinerU2.5-Pro0.8206
3Dots.mocr0.8199
4Chandra-OCR-20.8182
5Infinity-Parser2-Pro0.7852
6Jina-OCR-v10.7808
7TeleOCR0.7619
8Surya OCR 20.7619
9GLM-OCR0.7594
10PaddleOCR-VL-1.60.7559
11FireRed-OCR0.7477
12HunyuanOCR 1.50.7462
13LightOnOCR-2-1B0.738
14OvisOCR20.7348
15Youtu-Parsing0.7317

Interactive version: theaggregate.ai/benchmark?slug=docparsingbench-formula · How It Works · Data refreshed daily, snapshot 2026-11-16.