Unlimited-OCR

Name: Unlimited-OCR
Author: Baidu

Baidu · released 2026-06-22 · data checked 2026-11-06

Unlimited-OCR is a 3.3B-parameter OCR and document-parsing model from Baidu (CN), released 2026-06-22 under the MIT license. It scores 94.00 on OmniDocBench v1.6 (#13 of 33). It needs about 7.7 GB of memory at 16-bit.

Builds on DeepSeek-OCR for long documents.

OmniDocBench v1.6

94.00

#13 of 33 models

olmOCR-Bench

—

Not reported

Parameters

3.3B

OCR specialist

Memory to run

~7.7 GB

16-bit weights plus about 1 GB

Benchmarks

OmniDocBench v1.6 Leaderboard94.00

Best: OXR 1.0 97.03

The tick marks the best score on each benchmark. See how to read the ranking.

Specifications

Lab: Baidu (CN)
Type: OCR / document parsing specialist
Total parameters: 3.3B
License: MIT
Memory to run: ~7.7 GB at 16-bit
Released: 2026-06-22

Weights and usage on Hugging Face ↗

Closest alternatives

Infinity-Parser2 Pro (35.1B, 93.95 on OmniDocBench, Apache-2.0)
Qianfan-OCR (4.7B, 93.90 on OmniDocBench, Apache-2.0)
MiniCPM-V 4.7 35B-A3B (35.2B, 94.24 on OmniDocBench, Apache-2.0)
Youtu-Parsing (2.5B, 93.74 on OmniDocBench, Youtu-Parsing license)

Sources

Something wrong here? Report it.