Azure Document Intelligence vs Donut
Pick Azure Document Intelligence if you want a managed API that reads invoices, receipts, IDs, W-2s and contracts today, with per-page pricing and an on-prem container option. Pick Donut if you have a narrow, fixed-layout document type, labelled data to fine-tune on, and want an MIT-licensed model you run yourself. The main split is prebuilt and paid versus free but untrained.
Azure Document Intelligenceazure/document-intelligence | Donutnaver/donut | |
|---|---|---|
| Type | OCR API | Open source |
| Price | 60 credits / page | 1 credit / page |
| Per 1,000 pages | $60.00 | $1.00 |
| Typical speed | 1.5 s / page | 1.0 s / page |
| Best for | Enterprise forms and IDs | Receipt and form fields after fine-tuning |
| Inputs | PDF, images | images |
| Languages | 300+ languages | English, Chinese, Japanese, Korean |
| Max pages per request | 2,000 | 100 |
| Tables | ✓ Yes | No |
| Handwriting | ✓ Yes | No |
| Math | No | No |
| Bounding boxes | ✓ Yes | No |
| Confidence | ✓ Yes | No |
| Forms | ✓ Yes | ✓ Yes |
| 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 Azure Document Intelligence if
- It ships prebuilt models for invoices, receipts, IDs, W-2 and contracts, so you can extract fields without training anything.
- Read OCR costs $1.50 per 1,000 pages and drops to $0.60 above 1M pages, with 500 free pages per month.
- Disconnected containers let you run it on-prem if you are already an Azure shop with data residency needs.
Choose Donut if
- It is MIT licensed and free to use, so you only pay for your own compute.
- It skips the separate OCR step and maps the document image straight to structured JSON, which suits receipt and form field extraction after fine-tuning.
- A practitioner on Hacker News reports that about 200 documents of fine-tuning can reach human-level accuracy on narrow data.
Key differences
| Azure Document Intelligence | Donut | |
|---|---|---|
| Licence | Proprietary | MIT |
| How you run it | Managed Azure service, or containers including disconnected on-prem | Self-host only, available in Hugging Face Transformers |
| Price | Read $1.50 per 1,000 pages; Layout and prebuilt $10 per 1,000; custom extraction $30 per 1,000 | Free, you pay for compute |
| Ready out of the box | Yes: Read, Layout and prebuilt models for invoices, receipts, IDs, W-2 and contracts | No: the base checkpoint is pretrain-only and must be fine-tuned first |
| Output formats | JSON, Markdown, plain text, bounding boxes, searchable PDF | JSON, plain text |
| Benchmark | 82.7% on RD-TableBench table accuracy (independent) | Not published |
| Known failure | Costs multiply when models are stacked: Prebuilt plus Layout plus high-res add-on is $26 per 1,000 pages | Misses tiny text in large images because of input resolution limits; a Hacker News user found results on personal documents did not match the promised scores |
Sources
- Azure pricing page
- Azure DI pricing breakdown
- Reducto RD-TableBench
- donut-base model card
- Donut ECCV paper
- Hacker News thread
What each one is
Azure Document Intelligence
Azure layout model with tables, selection marks and prebuilt invoice and ID readers.
Azure Document Intelligence detailsDonut
OCR-free document understanding model. Fine-tuned for receipts, forms and document Q&A.
Donut 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": "azure/document-intelligence","fallbacks": ["naver/donut"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Azure Document Intelligence alternativesDonut alternativesBest OCR models in 2026