DeepSeek Unveils DeepSeek-OCR 2 Model
3 week ago / Read about 0 minute
Author:小编   

On January 27, 2026, the DeepSeek team released a research paper entitled 'DeepSeek-OCR 2: Visual Causal Flow' and, at the same time, made the DeepSeek-OCR 2 model available as open - source. This model adopts the groundbreaking DeepEncoder V2 approach. Through this approach, AI can dynamically reorganize different parts of an image according to the semantic information it contains, effectively mimicking human visual reasoning. When dealing with images that have intricate layouts, this model surpasses traditional visual - language models, delivering a more intelligent form of visual comprehension.