微软发布首个自主研发图像生成器MAI-Image-1

Core Insights - Microsoft has launched its first self-designed and developed text-to-image model, MAI-Image-1, marking a significant step in its AI journey [1] - The company has incorporated feedback from professional creatives during the development process to avoid "repetitive or generic outputs" [1] - MAI-Image-1 is reported to excel in generating realistic images such as lightning and landscapes, and it processes requests faster than "larger, slower models" [1] - The model has ranked among the top ten in the international large model competition arena, LMArena [1]