As reported by The Verge, Microsoft has successfully integrated its first in-house developed AI image generation model, MAI-Image-1, into two of its offerings: Bing Image Creator and Copilot Audio Expressions (the latter has not been launched in the EU region yet). This model made its official debut in October 2025. Mustafa Suleyman, who is at the helm of Microsoft AI, highlighted that MAI-Image-1 truly shines when it comes to generating images of food, natural landscapes, and creating realistic lighting effects. It boasts remarkable strengths in handling reflections, including diffuse reflections. Moreover, it strikes an excellent balance between generation speed and image quality. At present, within Bing's image generation service, MAI-Image-1 stands as one of the three available models for users to choose from, the other two being DALL-E 3 and GPT-4o.
