MinerU vs Pixtral
Pick MinerU for batch parsing of academic papers and mixed PDF and Office files, especially when you need Markdown, LaTeX or DOCX out and want to run on your own hardware. Pixtral 12B suits question answering over charts and documents, and only where you are already on the Mistral API. Mistral marks Pixtral as deprecated, so do not start a new OCR project on it.
MinerUopendatalab/mineru | Pixtralmistral/pixtral | |
|---|---|---|
| Type | Open source | Vision model |
| Price | 2 credits / page | 2 credits / page |
| Per 1,000 pages | $2.00 | $2.00 |
| Typical speed | 2.0 s / page | 3.0 s / page |
| Best for | Academic PDFs | Questions about document images |
| Inputs | PDF, images | images |
| Languages | 80+ languages | Multilingual |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | No |
| Handwriting | No | No |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | No |
| Confidence | No | No |
| Forms | No | 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 MinerU if
- It takes PDF, images, DOCX, PPTX, XLSX, EPUB, OFD and HTML in one tool and writes Markdown, HTML, LaTeX, DOCX, EPUB or JSON.
- The basic tier runs on CPU with about 2 GB RAM, and the VLM tier reaches 2.12 pages per second on one A100 with vLLM.
- It scores 75.2 on olmOCR-Bench (version 2.5.4), a score from an independent benchmark.
Choose Pixtral if
- It scores 90.7 ANLS on DocVQA, as reported by Mistral, which fits question answering over documents and charts.
- The Mistral API launch price was $0.15 per million input and output tokens, so teams already on that API can keep their integration.
- Its 128K token context allows many images in one request, and the weights are Apache-2.0.
Key differences
| MinerU | Pixtral | |
|---|---|---|
| Licence | MinerU Open Source License (Apache-2.0 based with extra conditions); MinerU2.5 VLM weights tagged AGPL-3.0 | Apache-2.0 |
| How you run it | Self-host only: CLI, API or WebUI, from CPU-only to VLM tiers | Mistral API (pixtral-12b-2409) or self-host the open weights |
| Hardware | Basic tier: CPU with 2 GB RAM. VLM tiers: 16 GB RAM and a GPU with 8 GB+ VRAM | Not published |
| Model size | MinerU2.5 VLM: 1.2B parameters | 12.4B parameters (12B decoder + 400M vision encoder) |
| Output formats | Markdown, HTML, LaTeX, DOCX, EPUB, JSON | Plain text, Markdown, JSON |
| Benchmarks | olmOCR-Bench 75.2 (independent); ParseBench mean 45.9 (independent) | DocVQA 90.7 ANLS, ChartQA 81.8, MMMU 52.5 (all from Mistral) |
| Known failures | Repeated content (300 KB figure blocks), dropped merged table header rows such as bank statement account lines, full-width Chinese punctuation turned half-width | Does not always get OCR right per a launch thread user; reported drops on handwriting, small fonts and non-Latin scripts; Q4 quantization degrades small text |
Sources
- MinerU GitHub README
- olmOCR-Bench table
- Issue #5554 table header rows
- Mistral announcement
- Mistral Sept 2024 pricing
- Third-party deployment guide
What each one is
MinerU
Open-source PDF to Markdown pipeline popular for academic and Chinese documents.
MinerU detailsPixtral
Mistral's open-weight vision models, Pixtral 12B and Pixtral Large. Mistral's API no longer serves them, so they run self-hosted.
Pixtral 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": "opendatalab/mineru","fallbacks": ["mistral/pixtral"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
MinerU alternativesPixtral alternativesBest OCR models in 2026