PaddleOCR-VL vs Pixtral
Pick PaddleOCR-VL for new OCR work: multilingual scans, tables, formulas and charts parsed into Markdown or JSON on a small self-hosted GPU. Pick Pixtral only if you already call it through the Mistral API and need question answering over charts and documents. Mistral marks Pixtral 12B as deprecated and no longer maintained, so it is a poor base for a new project.
PaddleOCR-VLpaddle/paddleocr-vl | Pixtralmistral/pixtral | |
|---|---|---|
| Type | Open source | Vision model |
| Price | 6 credits / page | 2 credits / page |
| Per 1,000 pages | $6.00 | $2.00 |
| Typical speed | 1.3 s / page | 3.0 s / page |
| Best for | CJK documents | Questions about document images |
| Inputs | PDF, images | images |
| Languages | 109 languages | Multilingual |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | No |
| Handwriting | No | No |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | No |
| Confidence | ✓ Yes | 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 PaddleOCR-VL if
- It supports 109 languages in one model, including Cyrillic, Arabic, Devanagari and Thai scripts.
- At 0.9B parameters it runs on much smaller GPUs than a 12.4B model, and it recognizes text, tables, formulas and charts as Markdown and JSON.
- Version 1.6 scores 96.33 on OmniDocBench v1.6, and the base model scores 80.0 on olmOCR-Bench in an independent table.
Choose Pixtral if
- It is a general vision language model that answers questions about charts and documents, with 90.7 ANLS on DocVQA and 81.8 on ChartQA, per Mistral.
- You can use it through the Mistral API at a launch price of $0.15 per million input and output tokens, so there is no GPU to run.
- Its 128K token context allows many images in one request, which suits existing Mistral API integrations.
Key differences
| PaddleOCR-VL | Pixtral | |
|---|---|---|
| Purpose | Document parser for text, tables, formulas and charts | General vision language model, not built only for OCR |
| Licence | Apache-2.0 | Apache-2.0 open weights |
| How you run it | Self-host only, via PaddleOCR, vLLM or transformers | Mistral API (pixtral-12b-2409) or self-host the open weights |
| Size and hardware | 0.9B parameters; NVIDIA GPU recommended, CPU possible but slow, no minimum VRAM stated | 12.4B parameters (12B decoder + 400M vision encoder); hardware needs not published |
| Price | Free, you pay for compute | $0.15 per million input and output tokens at API launch, or free to self-host |
| Benchmarks | OmniDocBench v1.6 96.33 (v1.6), olmOCR-Bench 80.0 | DocVQA 90.7 ANLS, ChartQA 81.8, MMMU 52.5 |
| Known failure | Version 1.6 regresses Russian and Kazakh Cyrillic on born-digital PDFs, and hangs on blurry images | Reported drops on handwriting, small fonts and non-Latin scripts; Q4 quantization degrades small text; deprecated |
Sources
- PaddleOCR-VL-1.6 model card
- olmOCR-Bench table
- Issue #18088 Cyrillic regression
- Mistral announcement
- Mistral Sept 2024 pricing
- Third-party deployment guide
What each one is
PaddleOCR-VL
Open-source document model from Baidu. Excellent on Chinese, Japanese and Korean.
PaddleOCR-VL 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": "paddle/paddleocr-vl","fallbacks": ["mistral/pixtral"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
PaddleOCR-VL alternativesPixtral alternativesBest OCR models in 2026