WritingBaidu (ERNIE)Baidu (ERNIE)published Oct 16, 2025seen Jun 5PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language ModelOpen original ↗Captured sourcesource ↗/b/blog/posts/paddleocr-vl///blog/posts/paddleocr-vl/PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language ModelSource ↗published Oct 16, 2025seen Jun 5No source text has been captured for this signal yet. The original source is linked below.source ↗Notabilitynotability 6.0/10Notable model release from Baidu, specialized but not frontier.