Nanonets-OCR vs InternVL3.5
Pick Nanonets-OCR if you self-host and want printed business PDFs turned into Markdown, HTML or LaTeX, with signatures, watermarks, checkboxes and flowcharts (as Mermaid) tagged. Pick InternVL3.5 if you want Apache-2.0 weights, a choice of sizes from 1B to 241B, and a general model for document question answering. The main split is scope: Nanonets-OCR is a fixed-prompt document converter, and InternVL3.5 is a general vision language model with no published OCR scores.
Nanonets-OCRnanonets/nanonets-ocr | InternVL3.5opengvlab/internvl | |
|---|---|---|
| Type | Open source | Vision model |
| Price | 2 credits / page | 2 credits / page |
| Per 1,000 pages | $2.00 | $2.00 |
| Typical speed | 1.8 s / page | 3.0 s / page |
| Best for | Markdown with semantic tags | Self-hosted document understanding |
| Inputs | PDF, images | images |
| Languages | Multilingual | Multilingual |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | No | No |
| Math | ✓ Yes | ✓ Yes |
| 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 converts checkboxes to unicode symbols and tags signatures, watermarks and image descriptions, which suits forms and business documents.
- It extracts flowcharts and org charts as Mermaid code and tells inline LaTeX equations from display ones.
- It publishes numbers you can check: DocVQA 89.43, ChartQA 78.56 and olmOCR-bench 69.5 for the OCR2-3B model.
Choose InternVL3.5 if
- Its Apache-2.0 licence is stated, while the Nanonets-OCR2-3B model card carries no licence tag.
- Nine sizes from 1B to 241B-A28B let you trade accuracy for cost, and models up to 30B fit on a single A100.
- It is a general model that answers questions about a document and is not tied to one fixed prompt.
Key differences
| Nanonets-OCR | InternVL3.5 | |
|---|---|---|
| Licence | Not stated on the Nanonets-OCR2-3B model card; base model is Qwen2.5-VL-3B | Apache-2.0 |
| Model size | 3B (Hugging Face lists about 4B) | 1B to 241B-A28B; the 8B variant is 8.5B |
| Output formats | Markdown, HTML, LaTeX, Mermaid | Plain text, Markdown, JSON |
| Published benchmarks | olmOCR-bench 69.5, DocVQA 89.43, ChartQA 78.56 (OCR2-3B, from the authors) | Not published as text values; the model card shows OCR scores only as chart images |
| Prompting | Trained on a fixed prompt; the team says hallucination with other prompts is expected | General VLM; no fixed-prompt limit is listed |
| Handwriting | Not trained on handwriting and can hallucinate; one test saw generation loops with CER up to 2100% and 69.1s per image on an L4 | Not published |
| Layout and coordinates | Not published | No bounding boxes or layout output by default, so not suited to word coordinates |
Sources
- Nanonets-OCR2-3B model card
- Nanonets-OCR-s announcement
- HF discussion fixed prompt
- Japanese handwriting OCR comparison
- Hugging Face model card (8B)
- InternVL GitHub
What each one is
Nanonets-OCR
Open image-to-Markdown model that tags tables, equations, signatures, watermarks and checkboxes.
Nanonets-OCR detailsInternVL3.5
Shanghai AI Lab's open vision-language family (1B to 241B, August 2025), strong on OCR and document benchmarks.
InternVL3.5 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": ["opengvlab/internvl"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Nanonets-OCR alternativesInternVL3.5 alternativesBest OCR models in 2026