Leon5201314 (Yuliang)

0.7B MonkeyOCRv2 Outperforms Larger Models on 17-Language Document Parsing

MonkeyOCRv2-B-Parsing reaches 83.3 on MDPBench, a multilingual benchmark covering digital-born and photographed documents across 17 languages.

Results among evaluated open-source models:

• MonkeyOCRv2-B, 0.7B: 83.3
• dots.mocr, 3B: 80.5
• HunyuanOCR-1.5, 1B: 76.8
• PaddleOCR-VL-1.6, 0.9B: 75.0
• MinerU2.5-Pro, 1.2B: 71.0

The central idea is simple: before asking an LLM to reason over a document, the vision encoder must preserve every character stroke, digit, punctuation mark, and layout cue.

MonkeyOCRv2 is pretrained on 113M document images across 17 languages using joint image-to-text generation and pixel-level reconstruction.

Models:
https://huggingface.co/collections/zenosai/monkeyocrv2

Paper:
MonkeyOCRv2: A Visual-Text Foundation Model for Document AI (2607.11562)

GitHub:
https://github.com/Yuliang-Liu/MonkeyOCRv2

Code and model weights are available under Apache-2.0.

We welcome tests on difficult multilingual, photographed, and visually ambiguous documents—especially failure cases.

1 reply

·

liked a Space about 1 month ago

MDPBench Leaderboard

📄

8

Explore multilingual document parsing leaderboard

liked a dataset about 1 month ago

zenosai/MonkeyDocv2

Updated Jul 21 • 27 • 2

liked a model about 1 month ago

zenosai/MonkeyOCRv2-S

Image Feature Extraction • 28.5M • Updated about 1 month ago • 1.1k • 13

posted an update about 1 month ago

Post

1299

0.7B MonkeyOCRv2 Outperforms Larger Models on 17-Language Document Parsing

MonkeyOCRv2-B-Parsing reaches 83.3 on MDPBench, a multilingual benchmark covering digital-born and photographed documents across 17 languages.

Results among evaluated open-source models:

• MonkeyOCRv2-B, 0.7B: 83.3
• dots.mocr, 3B: 80.5
• HunyuanOCR-1.5, 1B: 76.8
• PaddleOCR-VL-1.6, 0.9B: 75.0
• MinerU2.5-Pro, 1.2B: 71.0

The central idea is simple: before asking an LLM to reason over a document, the vision encoder must preserve every character stroke, digit, punctuation mark, and layout cue.

MonkeyOCRv2 is pretrained on 113M document images across 17 languages using joint image-to-text generation and pixel-level reconstruction.

Models:
https://huggingface.co/collections/zenosai/monkeyocrv2

Paper:
MonkeyOCRv2: A Visual-Text Foundation Model for Document AI (2607.11562)

GitHub:
https://github.com/Yuliang-Liu/MonkeyOCRv2

Code and model weights are available under Apache-2.0.

We welcome tests on difficult multilingual, photographed, and visually ambiguous documents—especially failure cases.