olmOCR vs Hunyuan OCR
Pick olmOCR for bulk PDF to Markdown, especially academic papers with equations and pages mixing print and handwriting. It scores 82.4 on olmOCR-Bench and costs under $200 per million pages self-hosted. Pick Hunyuan OCR for photo and scene text, image translation, or low-resource scripts, and when you need a ~1B model that runs on a laptop via llama.cpp.
olmOCRallenai/olmocr | 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 | 1.5 s / page | 1.5 s / page |
| Best for | Bulk PDF to text for datasets | Chinese documents and photos |
| Inputs | PDF, images | PDF, images |
| Languages | Mainly English | 100+ languages |
| Max pages per request | 1,000 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | No |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | No | ✓ Yes |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | No | ✓ 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 olmOCR if
- It scores 82.4 on olmOCR-Bench, above Marker (76.1), MinerU 2.5.4 (75.2) and Mistral OCR (72.0) in the same table.
- The toolkit is Apache-2.0 and the weights, training data and code are all released.
- It handles equations, tables and handwriting, and strips headers and footers in natural reading order.
Choose Hunyuan OCR if
- One ~1B model covers document parsing, text spotting, information extraction and image translation.
- It runs on CPU, consumer GPUs or laptops through the llama.cpp GGUF path, and it supports images up to 4K.
- It exposes 12 task types via the vLLM client, including low-resource and ancient-script OCR.
Key differences
| olmOCR | Hunyuan OCR | |
|---|---|---|
| Licence | Apache-2.0 (toolkit) | Tencent Hunyuan Community License; excludes EU, UK and South Korea; separate license above 100M monthly active users |
| Model size | 7B | ~1B |
| Hardware | NVIDIA GPU with at least 12 GB VRAM and 30 GB free disk; no CPU-only use | NVIDIA GPU with CUDA 13 recommended for the vLLM environment; llama.cpp GGUF for CPU, consumer GPU or laptop |
| How you run it | Self-host with vLLM, or use third-party hosted endpoints | Self-host only, via vLLM, Transformers or llama.cpp |
| Price | Under $200 per million pages self-hosted (Ai2); DeepInfra hosted olmOCR-2-7B at $0.09 input / $0.19 output per million tokens | Free, self-hosted (you pay for compute) |
| Output formats | Markdown, JSON | Markdown, HTML, LaTeX, JSON, plain text, bounding boxes |
| Benchmark | olmOCR-Bench overall 82.4 (v0.4.0), reported by the authors | ParseBench-2 overall 67.11 and ParseBench mean 63.86 (v1.5), independent; text score 62.88 |
| Known failure | Out of memory on 12 GB GPUs, can get stuck on the last few pages, unreliable image bounding box coordinates | Archived 1.0 note said Transformers and vLLM outputs differed in accuracy; unified environment needs CUDA 13 and vLLM 0.25.1+ |
Sources
What each one is
olmOCR
Open-source vision model from Ai2 (often searched as olm ocr) that turns PDFs into clean, reading-order text. Latest release: olmOCR 2.
olmOCR 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": "allenai/olmocr","fallbacks": ["tencent/hunyuan-ocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
olmOCR alternativesHunyuan OCR alternativesBest OCR models in 2026