MonkeyOCR vs Florence-2
Pick MonkeyOCR to convert English or Simplified Chinese PDFs, such as academic papers with formulas and tables, into structured output. Pick Florence-2 for short text in images, such as scene text and labels, where you need boxes and a small MIT-licensed model. MonkeyOCR returns document structure but its v1 weights need a written license for commercial use. Florence-2 returns only text and boxes.
MonkeyOCRmonkeyocr/monkeyocr | Florence-2microsoft/florence-2 | |
|---|---|---|
| Type | Open source | Open source |
| Price | 2 credits / page | 1 credit / page |
| Per 1,000 pages | $2.00 | $1.00 |
| Typical speed | 1.5 s / page | 0.8 s / page |
| Best for | Chinese and English PDFs | Lightweight OCR with boxes |
| Inputs | PDF, images | images |
| Languages | Chinese, English | English |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | No |
| Handwriting | No | No |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | No | No |
| 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 MonkeyOCR if
- Its pro-3B model reports an OmniDocBench overall edit distance of 0.138 in English and 0.206 in Chinese, and a table TEDS of 81.5 in English and 87.5 in Chinese.
- It outputs Markdown, JSON, LaTeX, HTML and bounding boxes, so tables and formulas survive conversion.
- It runs on an 8GB RTX 4060 with quantized weights, and pro-1.2B is about 36% faster than pro-3B for about 1.6% lower accuracy.
Choose Florence-2 if
- It is MIT licensed, so you can use it commercially without asking anyone.
- It is very small, at 0.23B for base and 0.77B for large, which suits lightweight edge OCR in English.
- OCR_WITH_REGION returns quadrilateral boxes per text line, and the same model also does captioning, detection and grounding.
Key differences
| MonkeyOCR | Florence-2 | |
|---|---|---|
| Licence | Code Apache-2.0; v1 weights are research and non-commercial only, commercial use needs a written license | MIT |
| Model size | 3B and 1.2B | 0.23B (base), 0.77B (large) |
| Output formats | Markdown, JSON, LaTeX, HTML, bounding boxes | Plain text, bounding boxes |
| Document structure | Parses full documents with formulas and tables | No Markdown, tables or LaTeX, just text and boxes |
| Chinese | Reports Chinese results on OmniDocBench for Simplified Chinese; Traditional Chinese is not fully supported | Chinese output is mostly wrong because the vocabulary lacks CJK characters |
| Handwriting | Not fully supported, along with photographed text | Spelling errors and missing spaces on cursive handwriting and French text in one blog test |
| Speed and setup | About 0.49 pages/s on a 3090 for pro-3B; Docker on 20/30/40-series and V100 needs a patched image | Not published |
Sources
What each one is
MonkeyOCR
Open-source document parser that splits work into structure, recognition and relation steps.
MonkeyOCR detailsFlorence-2
Small Microsoft vision model (0.23B and 0.77B) with OCR and OCR-with-regions tasks that return text with boxes.
Florence-2 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": "monkeyocr/monkeyocr","fallbacks": ["microsoft/florence-2"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
MonkeyOCR alternativesFlorence-2 alternativesBest OCR models in 2026