Nanonets-OCR vs Typhoon OCR
Pick Nanonets-OCR for printed business documents that need signatures, watermarks, checkboxes or flowcharts turned into Markdown, HTML, LaTeX or Mermaid, and you are happy to self-host. Pick Typhoon OCR for Thai and English documents such as government forms, financial reports and books, or when you want a hosted API as well as open weights. The main split is language coverage and benchmark strength: Typhoon V2 2B scores 84.7 on olmOCR-bench against 69.5 for Nanonets-OCR2-3B.
Nanonets-OCRnanonets/nanonets-ocr | Typhoon OCRtyphoon/typhoon-ocr | |
|---|---|---|
| Type | Open source | Open source |
| Price | 2 credits / page | 2 credits / page |
| Per 1,000 pages | $2.00 | $2.00 |
| Typical speed | 1.8 s / page | 1.8 s / page |
| Best for | Markdown with semantic tags | Thai documents |
| Inputs | PDF, images | PDF, images |
| Languages | Multilingual | Thai, English |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | No | No |
| Math | ✓ Yes | No |
| Bounding boxes | No | No |
| Confidence | No | No |
| Forms | ✓ Yes | 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 Nanonets-OCR if
- It tags signatures, watermarks and image descriptions, and converts checkboxes to unicode symbols, which suits business forms.
- It extracts flowcharts and org charts as Mermaid code, and it separates inline from display LaTeX equations.
- Its model card reports DocVQA 89.43 and ChartQA 78.56 for the OCR2-3B model.
Choose Typhoon OCR if
- It is built for Thai and English. V2 2B scores 81.3 on MDPBench Thai pages, against 61.3 for its base model.
- V1.5 beat GPT-5 and Gemini 2.5 Pro on Thai government forms (BLEU 0.870), and tables come out as HTML with merged cells.
- It scores 84.7 on olmOCR-bench at 2B parameters, and you can run it self-hosted or through the Typhoon API as typhoon-ocr-v2.
Key differences
| Nanonets-OCR | Typhoon OCR | |
|---|---|---|
| Deployment | Self-host only. Weights from Hugging Face. | Self-host, or the Typhoon API as typhoon-ocr-v2. |
| Licence | Not stated on the Nanonets-OCR2-3B model card (no license tag). Base model is Qwen2.5-VL-3B. | Apache-2.0, subject to OpenTyphoon Terms and Conditions. |
| Size | 3B (Hugging Face lists about 4B params). | 2B for V2 and V1.5, plus a 0.8B V2 variant. |
| Output formats | Markdown, HTML, LaTeX, Mermaid. | Markdown, HTML, LaTeX. |
| olmOCR-bench | 69.5 (OCR2-3B), reported by the authors. | 84.7 (V2 2B), reported by the authors. |
| Languages and handwriting | No language list published. Not trained on handwriting, and a handwriting test saw generation loops with CER up to 2100%. | Only Thai and English evaluated. Handwriting: Not published. |
| Known failure | Needs its fixed training prompt. The team says hallucination with other prompts is expected. | V2 can loop until the token limit on dense multi-column pages. Key-information extraction scores only 21.0 on ThaiOCRBench. |
Sources
- Nanonets-OCR2-3B model card
- Nanonets-OCR-s announcement
- HF discussion fixed prompt
- Typhoon OCR V2 2B model card
- Typhoon OCR technical report
What each one is
Nanonets-OCR
Open image-to-Markdown model that tags tables, equations, signatures, watermarks and checkboxes.
Nanonets-OCR detailsTyphoon OCR
Open vision model tuned for Thai and English documents, with Markdown output.
Typhoon 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": "nanonets/nanonets-ocr","fallbacks": ["typhoon/typhoon-ocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Nanonets-OCR alternativesTyphoon OCR alternativesBest OCR models in 2026