Typhoon OCR vs Florence-2
Pick Typhoon OCR for full-page Thai and English documents where you need Markdown, HTML tables or LaTeX out. Pick Florence-2 when you need a very small English model that returns text with bounding boxes, or one model that also does captioning and detection. The main split is structure: Typhoon OCR rebuilds the page, Florence-2 returns only text and boxes.
Typhoon OCRtyphoon/typhoon-ocr | 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 | 1.8 s / page | 0.8 s / page |
| Best for | Thai documents | Lightweight OCR with boxes |
| Inputs | PDF, images | images |
| Languages | Thai, English | English |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | No |
| Handwriting | No | No |
| Math | No | No |
| Bounding boxes | No | ✓ Yes |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | ✓ Yes | 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 Typhoon OCR if
- It is built for Thai government forms, financial reports, Thai books and mixed Thai-English documents.
- It outputs Markdown, HTML and LaTeX, with tables as HTML with merged cells and checkboxes marked.
- It scores 84.7 on olmOCR-bench at 2B parameters, and the V1.5 model reached BLEU 0.870 on Thai government forms.
Choose Florence-2 if
- At 0.23B (base) or 0.77B (large) and MIT licensed, it suits lightweight edge OCR in English.
- OCR_WITH_REGION returns quadrilateral boxes for each text line, which suits scene text and short labels.
- The same model also does captioning, detection and grounding, so one model can cover a pipeline that needs more than OCR.
Key differences
| Typhoon OCR | Florence-2 | |
|---|---|---|
| Output formats | Markdown, HTML, LaTeX | Plain text and bounding boxes. No Markdown, tables or LaTeX. |
| Licence | Apache-2.0, subject to OpenTyphoon Terms and Conditions | MIT |
| How you run it | Self-host the weights, or use the Typhoon API as typhoon-ocr-v2 | Self-host via Transformers only |
| Model size | 2B (V2 and V1.5), plus a 0.8B V2 variant | 0.23B (base), 0.77B (large) |
| Price | Free self-hosted, you pay for compute. API pricing is not published on the model card. | Free self-hosted, you pay for compute |
| Languages | Thai and English evaluated. No other languages tested. | English is the fit. Chinese output is mostly wrong because the vocabulary lacks CJK characters. |
| Known failure | V2 can loop until the token limit on dense multi-column pages. Key-information extraction is weakest, at 21.0 on ThaiOCRBench. | Spelling errors and missing spaces on cursive handwriting and French text in a Hugging Face blog test. |
Sources
- Typhoon OCR V2 2B model card
- Typhoon OCR technical report
- Florence-2-large model card
- HF blog OCR test
- HF discussion Chinese very bad
What each one is
Typhoon OCR
Open vision model tuned for Thai and English documents, with Markdown output.
Typhoon OCR 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": "typhoon/typhoon-ocr","fallbacks": ["microsoft/florence-2"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Typhoon OCR alternativesFlorence-2 alternativesBest OCR models in 2026