GPT vs OCRmyPDF
Pick GPT when you need data out of a document: handwritten notes, custom JSON fields, or question answering over a PDF. Pick OCRmyPDF when you need to make a pile of scanned PDFs searchable and archive-ready (PDF/A) on your own CPUs at no licence cost. GPT reads and extracts, while OCRmyPDF only adds a text layer.
GPTopenai/gpt | OCRmyPDFocrmypdf/ocrmypdf | |
|---|---|---|
| Type | Vision model | Open source |
| Price | 11 credits / page | 1 credit / page |
| Per 1,000 pages | $11.00 | $1.00 |
| Typical speed | 2.4 s / page | 1.0 s / page |
| Best for | Messy scans, custom instructions | Searchable PDFs from scans |
| Inputs | PDF, images | PDF, images |
| Languages | 50+ languages | 100+ languages (via Tesseract) |
| Max pages per request | 100 | 1,000 |
| Tables | ✓ Yes | No |
| Handwriting | ✓ Yes | No |
| Math | ✓ Yes | No |
| Bounding boxes | No | 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 GPT if
- You need handwritten notes read; one HN user reported about 5% errors on messy cursive.
- You want structured output such as Markdown or custom JSON from a page, not just text.
- You want to ask questions about a document or extract fields with reasoning, using files up to 50 MB per request.
Choose OCRmyPDF if
- You need validated searchable PDF/A output for archiving, with the OCR layer inserted losslessly where possible.
- You want a free, self-hosted tool under MPL-2.0 that runs CPU only and spreads work across all cores.
- You have scanned documents in many languages, since Tesseract language packs cover 100+ languages.
Key differences
| GPT | OCRmyPDF | |
|---|---|---|
| How you run it | Hosted API from OpenAI | Local command-line tool and library |
| Licence | Proprietary | MPL-2.0 (open source) |
| Price | Token-based per model. Page images add input tokens, so cost per page is not fixed. | Free and self-hosted. You pay for compute. |
| Output formats | Markdown, JSON, plain text | Searchable PDF/A, plain text |
| Handwriting | Listed as a good fit for handwritten notes | Listed as not suited to handwriting |
| Benchmark | OmniDocBench v1.5 overall 85.50 for GPT-5.2, against 90.33 for Gemini 3 Pro in the same paper | Not published |
| Known failures | No bounding boxes or confidence scores in the standard response | Output size can double or quadruple with --deskew --clean-final. --skip-text with default PDF/A output removes form fields. It hangs on PDFs whose forms share one Resources dictionary. |
Sources
- OpenAI file inputs docs
- PaddleOCR-VL-1.5 paper (GPT-5.2 eval)
- HN Mistral OCR thread
- OCRmyPDF GitHub README
- Issue #948 file size
- Issue #1790 form fields
What each one is
OCRmyPDF
Command-line tool that adds a searchable text layer to scanned PDFs, using Tesseract.
OCRmyPDF 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": ["ocrmypdf/ocrmypdf"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
GPT vs RapidOCRGPT vs EasyOCRClaude vs OCRmyPDFGemini Flash vs OCRmyPDFOCRmyPDF vs Qwen-VLOCRmyPDF vs dots.ocrGPT vs Amazon TextractMistral OCR vs OCRmyPDF
GPT alternativesOCRmyPDF alternativesBest OCR models in 2026