Woofun AI reports that DeepSeek has released its multimodal visual understanding model, V4-Flash-Vision-Exp, alongside public API access. Built upon the V4-Flash foundation, this iteration introduces capabilities for image comprehension and visual generation. The system supports multimodal input and output using both text and images, while providing developers with interfaces for inference and generation functions.