GLM-OCR vs Pixtral
Pick GLM-OCR for new OCR work: high-volume PDF to Markdown, academic papers with formulas, and schema-based field extraction. Pick Pixtral only if you already run on the Mistral API and need question answering over charts and documents, because Mistral marks it deprecated and unmaintained. Do not use either for financial tables where every row and decimal must survive.
GLM-OCRzai/glm-ocr | 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 | Low-cost general documents | Questions about document images |
| Inputs | PDF, images | images |
| Languages | 100+ 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 GLM-OCR if
- It is a 0.9B model built for OCR, and it scores 94.62 on OmniDocBench v1.5 per its own model card.
- The Z.ai API costs $0.03 per million tokens for input and output, and the MIT-licensed weights self-host with vLLM, SGLang, Ollama or transformers.
- It can extract fields into a user-defined JSON schema, so it fits forms as well as plain transcription.
Choose Pixtral if
- It reports 90.7 ANLS on DocVQA and 81.8 on ChartQA, which suits question answering over documents and charts.
- Its 128K token context allows many images in one request.
- It ships as Apache-2.0 open weights and is also available on the Mistral API, which helps if you already have that integration.
Key differences
| GLM-OCR | Pixtral | |
|---|---|---|
| Model size and type | 0.9B OCR model: PP-DocLayout-V3 layout analysis plus a GLM decoder | 12.4B general vision language model: 12B Mistral Nemo decoder plus 400M vision encoder |
| Licence | MIT for the weights; the SDK repo and the bundled PP-DocLayoutV3 are Apache-2.0 | Apache-2.0 |
| API price | $0.03 per million tokens, input and output | $0.15 per million input and output tokens at launch |
| Maintenance status | Active GitHub issue tracker; no deprecation notice | Mistral marks Pixtral 12B as deprecated and no longer maintained |
| Benchmarks | OmniDocBench v1.5 94.62 and olmOCR-Bench 75.2 (authors); ParseBench mean 29.6 (independent) | DocVQA 90.7 ANLS, ChartQA 81.8, MMMU 52.5 (authors) |
| Known failure modes | Repeats output unless a stop sequence is set; drops table footer and summary rows; loses comma decimal separators in cropped tables; no aarch64 Linux support | Does not always get OCR near correct per a launch thread user; reported accuracy drop on handwriting, small fonts and non-Latin scripts; Q4 quantization degrades small text OCR |
| Limits | API caps input at 100 pages, 50 MB per PDF and 10 MB per image | 128K token context; page and file size limits Not published |
Sources
- Hugging Face model card
- Z.ai GLM-OCR docs and pricing
- Issue #240 missing table rows
- Mistral announcement
- Mistral Sept 2024 pricing
- HN launch thread
What each one is
Pixtral
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.
PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?
Use one, fall back to the other
{"model": "zai/glm-ocr","fallbacks": ["mistral/pixtral"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
GLM-OCR vs InternVL3.5GLM-OCR vs RolmOCRUnlimited OCR vs PixtralPaddleOCR-VL vs PixtralGLM-OCR vs EasyOCRGLM-OCR vs Phi-4-multimodalEasyOCR vs PixtralDonut vs Pixtral
GLM-OCR alternativesPixtral alternativesBest OCR models in 2026