Tasks and inputs
MiniCPM-OCR 1.0 is listed for Image-to-text, Document OCR. It accepts image and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
MiniCPM-OCR 1.0 is OpenBMB's open 1.3B document-parsing model for text, tables, formulas and charts, with a free API and on-device apps.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: No API pricing stated; open weights for self-hosting
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
At 1.3B parameters, MiniCPM-OCR 1.0 is small enough to parse documents on a phone, and OpenBMB has shipped it there: iOS from 2 November 2026, Android and HarmonyOS from 12 November. That matters for anyone who cannot send scanned contracts or medical forms to a cloud API.
It is the first OCR-specific model in the MiniCPM-V/o family, built on MiniCPM-V 4.6, and was open-sourced on 27 October 2026 under Apache 2.0.
Text, tables, formulas, charts, seals and structured fields for key information extraction, in Chinese and English. It is tuned for real captures: scans, curved and skewed pages, photographs and documents re-photographed off a screen.
These are OpenBMB's figures.
| Route | Details |
|---|---|
| Free hosted API | Playground and API docs, China and global |
| Self-host | vLLM, SGLang, llama.cpp, Ollama; official Docker image since 5 November |
| Quantised | GGUF, AWQ, GPTQ; DSpark variants |
| SDK | minicpm-ocr on pip (1.0.5 adds multi-file runs) |
| Mobile | iOS, Android, HarmonyOS, offline |
| Agent | A one-prompt Skill that installs and runs it |
OpenBMB released MiniCPM-OCR Jev the same day: a 2B model that scores whether an OCR result is correct. Running both lets you auto-accept clean regions and send the doubtful ones to people.
Take a hundred documents from your actual backlog, including the worst phone photos. Compare table structure and formula accuracy against your current parser and log pages per second on your own hardware. Test the mobile build separately if on-device is the reason you are here. Browse alternatives in the document OCR directory and more from the lab on OpenBMB's provider page.
Compare the model on the work you need to complete.
MiniCPM-OCR 1.0 is listed for Image-to-text, Document OCR. It accepts image and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
Measure a completed task using the prompts, documents and tools your application will use. Include reasoning and retries in the result.
The weights are downloadable under Apache-2.0. Check runtime support and size the complete model and working context for your expected concurrency.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
OpenBMB
MiniCPM-OCR Jev is OpenBMB's open 2B judge that scores whether an OCR result for a text, formula or table region is correct, from 0 to 1.
Tencent
Youtu-Parsing-Omni is Tencent Youtu's 5B open parser that turns documents, charts, images, audio and video into one structured JSON format.
lightonai
LightOnOCR-3-0.8B is LightOn's smallest open document model: page images in, markdown out, with optional bounding boxes and chart tables.
Sarvam
Sarvam Vision 2.1 is Sarvam's 3B document-intelligence VLM for OCR, tables, forms and Indic handwriting in 22 Indian languages plus English.
NVIDIA
Nemotron 3.5 Parse is NVIDIA's 0.631B open document parser that turns page images into Markdown, HTML or JSON with tables, charts and reading order.
Cohere
Cohere Parse (parse-v5.0) is a 2.3B vision-language document parser that turns PDFs, slides and scans into Markdown for RAG, at $1.50 per 1,000 pages.