Nanonets-OCR vs Hunyuan OCR
Pick Nanonets-OCR for printed business documents where you need signatures, watermarks and checkboxes tagged, or flowcharts turned into Mermaid. Pick Hunyuan OCR for photo and scene text, translation of text in images, or low-resource and historical scripts, and when you want a ~1B model that runs on a laptop. Check the licence first: Hunyuan's does not apply in the EU, UK or South Korea, and Nanonets-OCR2-3B's model card states no licence.
Nanonets-OCRnanonets/nanonets-ocr | Hunyuan OCRtencent/hunyuan-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 | 1.5 s / page |
| Best for | Markdown with semantic tags | Chinese documents and photos |
| Inputs | PDF, images | PDF, images |
| Languages | Multilingual | 100+ languages |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | No | No |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | No | ✓ Yes |
| 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 turns checkboxes into unicode symbols, which suits forms and business documents.
- It extracts flowcharts and org charts as Mermaid code, and separates inline from display LaTeX equations.
- The 3B model scores 89.43 on DocVQA and 78.56 on ChartQA per its model card, with Markdown, HTML, LaTeX and Mermaid output.
Choose Hunyuan OCR if
- One ~1B model covers document parsing, text spotting, information extraction and image translation.
- It supports images up to 4K and a 128K context window in v1.5, and offers 12 task types including low-resource and ancient-script OCR.
- It runs on CPU, consumer GPUs or laptops through the llama.cpp GGUF path, and outputs JSON and bounding boxes as well as Markdown, HTML and LaTeX.
Key differences
| Nanonets-OCR | Hunyuan OCR | |
|---|---|---|
| Size | 3B (Hugging Face lists ~4B params), fine-tuned from Qwen2.5-VL-3B | ~1B |
| Licence | Not stated on the Nanonets-OCR2-3B model card (no licence tag); base model is Qwen2.5-VL-3B | Tencent Hunyuan Community License. Excludes the EU, UK and South Korea. A separate licence is needed above 100M monthly active users |
| How you run it | Self-host the weights from Hugging Face. Free, you pay for compute | Self-host via vLLM, Transformers or llama.cpp. Free, you pay for compute. CUDA 13 and vLLM 0.25.1+ for the unified vLLM environment |
| Output formats | Markdown, HTML, LaTeX, Mermaid | Markdown, HTML, LaTeX, JSON, plain text, bounding boxes |
| Benchmarks | olmOCR-bench overall 69.5, DocVQA 89.43, ChartQA 78.56 (all from the authors, OCR2-3B) | ParseBench-2 overall 67.11 and ParseBench mean 63.86 (independent, v1.5) |
| Prompts | Trained on a fixed prompt. The team says hallucination with other prompts is expected | 12 task types exposed through the vLLM client |
| Known failure | Not trained on handwriting and can hallucinate. In a handwriting test it showed generation loops, with CER up to 2100% and 69.1s per image on an L4 | Transformers and vLLM outputs differed in accuracy in the archived 1.0 note. ParseBench-2 text score is 62.88 in public CI replay |
Sources
- Nanonets-OCR2-3B model card
- Nanonets-OCR-s announcement
- HF discussion fixed prompt
- HunyuanOCR GitHub
- HunyuanOCR model card
- Hunyuan Community License
What each one is
Nanonets-OCR
Open image-to-Markdown model that tags tables, equations, signatures, watermarks and checkboxes.
Nanonets-OCR detailsHunyuan OCR
Small end-to-end OCR model from Tencent covering parsing, extraction and photo translation.
Hunyuan 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": ["tencent/hunyuan-ocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Nanonets-OCR alternativesHunyuan OCR alternativesBest OCR models in 2026