PaddleOCR (PP-OCR) vs dots.ocr
Pick PaddleOCR (PP-OCR) for fast text line detection and recognition on CPU, edge devices or in the browser, especially for Chinese and English receipts, signs and screenshots. Pick dots.ocr when you need whole documents turned into Markdown or JSON with layout, tables as HTML and formulas as LaTeX, and you have a CUDA GPU. The main split is scope: PP-OCR returns text lines and boxes, while dots.ocr parses full pages.
PaddleOCR (PP-OCR)paddle/paddleocr | dots.ocrrednote/dots-ocr | |
|---|---|---|
| Type | Open source | Open source |
| Price | 1 credit / page | 2 credits / page |
| Per 1,000 pages | $1.00 | $2.00 |
| Typical speed | 0.7 s / page | 1.5 s / page |
| Best for | Fast printed text, Chinese and English | Multilingual document parsing |
| Inputs | PDF, images | PDF, images |
| Languages | 80+ languages | 100+ languages |
| Max pages per request | 500 | 500 |
| Tables | No | ✓ Yes |
| Handwriting | No | No |
| Math | No | ✓ Yes |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | ✓ Yes | No |
| Forms | No | No |
| Multilingual | ✓ Yes | ✓ Yes |
| On OCRRouter | Available | Available |
Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.
Which to choose
Choose PaddleOCR (PP-OCR) if
- PP-OCRv6 medium is 34.5M parameters and runs on CPU, NVIDIA GPU, Kunlunxin XPU and other accelerators, with a 5.2x CPU speedup using OpenVINO per the README.
- It runs in the browser through the official PaddleOCR.js SDK, and also from Python or C++, which suits edge and mobile apps.
- One PP-OCRv6 model covers 50 languages, including Chinese, English, Japanese and 46 Latin-script languages.
Choose dots.ocr if
- One 3B model returns layout boxes, categories and text in a single pass, with tables as HTML and formulas as LaTeX.
- It outputs Markdown, JSON, HTML, LaTeX and bounding boxes, so multilingual PDFs, academic papers and business documents convert to Markdown with layout intact.
- The dots.mocr (v1.5) release scores 83.9 on olmOCR-bench, up from 79.1 for the original, and the dots.mocr-svg variant converts charts and diagrams to SVG.
Key differences
| PaddleOCR (PP-OCR) | dots.ocr | |
|---|---|---|
| What it returns | Text lines and bounding boxes; tables and formulas need the separate PP-Structure or VL pipelines | Layout, text, tables (HTML) and formulas (LaTeX) in one pass |
| Output formats | Plain text, bounding boxes, JSON | JSON, Markdown, HTML, LaTeX, bounding boxes, SVG |
| Model size | PP-OCRv6 medium: 34.5M parameters | 3B parameters (1.7B LLM backbone) |
| Hardware | CPU, NVIDIA GPU, Kunlunxin XPU; 0.13s on A100 reported for PP-OCRv6 | Single CUDA GPU with vLLM is the documented path; no VRAM figure stated; not suited to CPU-only production |
| Licence | Apache-2.0 | MIT, plus a separate dots.ocr license agreement file in the repository |
| Benchmark | Not published | olmOCR-bench 79.1 (dots.ocr), 83.9 (dots.mocr), per the authors |
| Known failure | PP-OCRv6 weights contain denormals that slow CPU inference 4 to 11x (RapidOCR issue #743) | Runs of ellipses or underscores can make output repeat endlessly; a table above 1000 grid cells drops the whole page in Docling (issue #5146) |
Sources
- PaddleOCR GitHub README
- PP-OCRv6 on Hugging Face
- RapidOCR issue #743 denormal weights
- dots.ocr GitHub README
- Docling issue #5146
What each one is
PaddleOCR (PP-OCR)
Baidu's classic open-source text detection and recognition pipeline. Light enough to run on CPU.
PaddleOCR (PP-OCR) detailsdots.ocr
Single vision model from rednote that does layout detection and content recognition together.
dots.ocr detailsTry both on your own file
Run the file once with each model and compare the output.
Use one, fall back to the other
{"model": "paddle/paddleocr","fallbacks": ["rednote/dots-ocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
PaddleOCR (PP-OCR) alternativesdots.ocr alternativesBest OCR models in 2026