OvisOCR2

OvisOCR2 Technical Report  |  Online Demo Introduction We are pleased to announce the release of OvisOCR2, a compact 0.8B end-to-end model for page-level document parsing. Given a document page image, OvisOCR2 generates a Markdown representation in natural reading order, covering text, formulas, tables, and visual regions. OvisOCR2 is developed by post-training Qwen3.5-0.8B using a carefully designed data engine that combines real-world and synthetic data, together with a multi-stage training recipe integrating SFT, RL, and OPD. The model delivers strong document parsing performance while maintaining a small deployment footprint. OvisOCR2 achieves an overall score of 96.58 on OmniDocBench v1.6, establishing a new state of the art and **becoming the first end-to-end model to top this leaderboard previously dominated by pipeline methods**. On PureDocBench, OvisOCR2 also achieves the highest Avg3 score of 75.06. Performance Inference By default, `parse` removes HTML image tags for visual regions. To render Markdown with visual regions, set `filter_imgtags=False` and save the Markdown file together with the referenced image crops as follows: Citation If you find OvisOCR2 useful, please consider citing our technical report: License This project is licensed under the Apache License, Version 2.0 (SPDX-License-Identifier: Apache-2.0). Disclaimer We used filtering and quality-assurance procedures during data construction to reduce parsing errors such as repeated outputs, incomplete content, invalid table/formula structures, and reading-order inconsistencies. Due to the diversity and complexity of real-world documents, OvisOCR2 may still produce incorrect or incomplete outputs. Please manually verify results in critical applications.