Hunyuan OCR vs RolmOCR
Pick RolmOCR for plain text transcription of scanned pages, especially rotated or skewed ones, under an Apache-2.0 licence on a single GPU. Pick Hunyuan OCR when you need Markdown, HTML, LaTeX, JSON or bounding boxes, or photo text spotting and image translation, and your users are outside the EU, UK and South Korea. The main split is output: RolmOCR returns plain text only, while Hunyuan OCR returns structured formats at about 1B parameters.
Hunyuan OCRtencent/hunyuan-ocr | RolmOCRreducto/rolmocr | |
|---|---|---|
| Type | Open source | Open source |
| Price | 2 credits / page | 2 credits / page |
| Per 1,000 pages | $2.00 | $2.00 |
| Typical speed | 1.5 s / page | 3.0 s / page |
| Best for | Chinese documents and photos | Fast open PDF to text |
| Inputs | PDF, images | PDF, images |
| Languages | 100+ languages | Not stated by the authors |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | No | ✓ Yes |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | No |
| 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 Hunyuan OCR if
- It outputs Markdown, HTML, LaTeX, JSON, plain text and bounding boxes, so you can get structure and coordinates that RolmOCR cannot give.
- One model of about 1B parameters covers document parsing, text spotting, information extraction and image translation, and it runs on laptops through llama.cpp GGUF.
- It handles images up to 4K with a 128K context window in v1.5, and offers 12 task types including low-resource and ancient-script OCR.
Choose RolmOCR if
- It is licensed Apache-2.0 and usable commercially, with no regional exclusions and no OSI-approval issue.
- About 15% of its training data was rotated, so it is built for off-angle and skewed scans.
- Needl.ai counted 7 hallucination issues for it versus 11 for olmocr-2-7b, and it fits self-hosted pipelines on one GPU with vLLM.
Key differences
| Hunyuan OCR | RolmOCR | |
|---|---|---|
| Licence | Tencent Hunyuan Community License; excludes EU, UK and South Korea; separate licence above 100M monthly active users | Apache-2.0 |
| Model size | About 1B parameters | 8B (Qwen2.5-VL-7B base) |
| Output formats | Markdown, HTML, LaTeX, JSON, plain text, bounding boxes | Plain text only; no layout bounding boxes |
| How you run it | Self-hosted with vLLM, Transformers or llama.cpp; unified vLLM environment needs CUDA 13 and vLLM 0.25.1+; GGUF path for CPU, consumer GPU or laptop | Self-hosted, for example with vLLM; quantized versions have not been evaluated |
| Price | Free, you pay for compute | Free, you pay for compute |
| Benchmark | ParseBench-2 overall 67.11; ParseBench mean 63.86 (v1.5) | Needl.ai PDF extraction overall 0.663, versus 0.770 for olmocr-2-7b |
| Known failure | Archived 1.0 note said Transformers and vLLM outputs differed in accuracy; text score 62.88 in public CI replay | Can hallucinate or drop content; omitted a table subtitle that olmOCR extracted |
Sources
What each one is
Hunyuan OCR
Small end-to-end OCR model from Tencent covering parsing, extraction and photo translation.
Hunyuan OCR detailsRolmOCR
Reducto's open PDF-to-text model, built on Qwen2.5-VL-7B with the olmOCR recipe. Needs no PDF metadata and handles rotated pages.
RolmOCR 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": "tencent/hunyuan-ocr","fallbacks": ["reducto/rolmocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Hunyuan OCR alternativesRolmOCR alternativesBest OCR models in 2026