GPT vs MinerU
Pick GPT when you need handwriting, custom JSON extraction or document question answering through a hosted API. Pick MinerU when you want to self-host, convert Chinese and English papers or mixed Office and PDF files, or run batch jobs on CPU. The main split is a paid, closed API versus a free, local toolkit with licence strings attached.
GPTopenai/gpt | MinerUopendatalab/mineru | |
|---|---|---|
| Type | Vision model | Open source |
| Price | 11 credits / page | 2 credits / page |
| Per 1,000 pages | $11.00 | $2.00 |
| Typical speed | 2.4 s / page | 2.0 s / page |
| Best for | Messy scans, custom instructions | Academic PDFs |
| Inputs | PDF, images | PDF, images |
| Languages | 50+ languages | 80+ languages |
| Max pages per request | 100 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | No |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | No | ✓ Yes |
| 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 GPT if
- It was reported at about 5% errors on messy cursive, against about 30% for Mistral OCR, so it suits handwritten notes.
- You can prompt it to extract straight into custom JSON, or to answer questions about a document.
- It is a hosted API that takes files up to 50 MB per request, so there is nothing to deploy.
Choose MinerU if
- It is free to self-host and the basic tier runs on CPU with about 2 GB RAM, so you pay only for compute.
- It reads PDF, images, DOCX, PPTX, XLSX, EPUB, OFD and HTML and renders to Markdown, HTML, LaTeX, DOCX, EPUB or JSON.
- The VLM tier reaches 2.12 pages per second on one A100 with vLLM, and it targets Chinese and English academic papers.
Key differences
| GPT | MinerU | |
|---|---|---|
| How you run it | Hosted OpenAI API | Self-host as CLI, API or WebUI |
| Licence | Proprietary | MinerU Open Source License (Apache-2.0 based with extra conditions); MinerU2.5 VLM weights tagged AGPL-3.0 |
| Price | Token-based; page images add input tokens, so cost per page is not fixed | Free software; you pay for compute |
| Output formats | Markdown, JSON, plain text | Markdown, HTML, LaTeX, DOCX, EPUB, JSON |
| Benchmark | OmniDocBench v1.5 overall 85.50 (GPT-5.2); v1.6 overall 86.59 | olmOCR-Bench overall 75.2 (MinerU 2.5.4); ParseBench mean 45.9 (MinerU2.5). Different benchmarks, so not directly comparable |
| Coordinates | No bounding boxes or confidence scores in the standard response | Not published |
| Known failure | Cost varies with page images; scored below Gemini 3 Pro (90.33) on OmniDocBench v1.5 | VLM tiers drop merged header rows in tables such as bank statements; repeated content can produce 300 KB figure blocks; full-width Chinese punctuation becomes half-width with no switch |
Sources
- OpenAI file inputs docs
- OmniDocBench leaderboard
- HN Mistral OCR thread
- MinerU GitHub README
- Issue #5554 table header rows
- olmOCR-Bench table
What each one is
MinerU
Open-source PDF to Markdown pipeline popular for academic and Chinese documents.
MinerU detailsTry both on your own file
Run the file once with each model and compare the output.
PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?
Use one, fall back to the other
{"model": "openai/gpt","fallbacks": ["opendatalab/mineru"],"file_url": "https://example.com/document.pdf"}