the AI bench
VERIFIED JULY 2026
All models

MODEL · BAIDU · COMPACT OCR VISION-LANGUAGE MODEL

Baidu Unlimited-OCR

Baidu's open-weight OCR / document-understanding VLM, positioned as a step beyond DeepSeek-OCR — high-fidelity text + layout extraction from images and documents. MIT-licensed, with vLLM support (community-added), an arXiv paper (2606.23050), and a live Hugging Face Space demo. Strong early traction (400K+ downloads within days).

License: MIT · Context: Document-page scale · Released: June 22, 2026

The decision in five lines

The call
Skip for local
Best for
Local evaluation and family reference
Runs on
No planner hardware fits at default quant — see model card
Watch out
General chat or reasoning — this is a document/OCR specialist, not a generalist VLM.
Evidence
Editorial · last verified July 2026

Compact OCR vision-language model
PARAMETERS
OCR / DOCUMENT VISION-LANGUAGE MODEL
TYPE
Document-page
CONTEXT
Runs on a single consumer GPU (compact VLM)
VRAM AT Q4

Where we recommend this

This model isn’t currently in an active planner slot. See the runner notes below if you’re running it anyway.

The call

Baidu's open-weight OCR / document-understanding VLM, positioned as a step beyond DeepSeek-OCR — high-fidelity text + layout extraction from images and documents. MIT-licensed, with vLLM support (community-added), an arXiv paper (2606.23050), and a live Hugging Face Space demo. Strong early traction (400K+ downloads within days).

When not to use: General chat or reasoning — this is a document/OCR specialist, not a generalist VLM. For broad vision-language work use MiniCPM-V-4.6 or MiniCPM-o 4.5.

Runner notes

vLLM inference supported (community PRs); also on ModelScope. GitHub `baidu/Unlimited-OCR` for the pipeline. MIT — clean for commercial document-processing use.

License
MIT
Released
June 22, 2026
Maker
Baidu

Next step

Find-by-model — see what hardware runs this