Baidu's PaddlePaddle team announced the release of PaddleOCR-VL, a new 0.9B-parameter vision-language model designed for end-to-end document parsing. The model is capable of processing complex, multilingual documents with various elements like dense layouts, different scripts, formulas, charts, and handwriting, converting them into structured formats like Markdown or JSON. PaddleOCR-VL supports 109 languages and reportedly achieves state-of-the-art performance on some benchmarks, outperforming larger models in page-level document parsing and element-level recognition. The release is seen as a significant step in making advanced document intelligence more efficient and accessible for practical deployment. No specific market reaction or analysis from major financial news outlets was available at the time of this report.