MAY-Image-1. This is the name of Microsoft’s first AI model capable of generating images. Completely developed in-house, “it excels at generating photorealistic images,” Microsoft said in a press release. MAI-Image 1 would be especially good for lighting (reflected light, reflections, etc.) or even landscapes.
Fast and with quality images, according to the American giant, it will allow users to quickly visualize their ideas on the screen, but also transfer them to other tools to improve them.
Reduce dependence on OpenAI
With MAI-Image 1, Microsoft aims to offer creators “real added value.” “This model is designed to provide real flexibility, visual diversity and practical value,” the company said. “MAI-Image-1 marks the next step in our journey and paves the way for more immersive, creative and dynamic experiences within our products,” he added.
But, with this AI model, Microsoft especially begins to reduce its dependence on OpenAI, a start-up with which the Redmond company has been associated since 2016. Thus, several of its tools are based on models from the creator of ChatGPT. Bing Image Creator, which, as its name suggests, uses AI to create images, works, for example, thanks to GPT-4o and Dall-E 3.
But that will change soon as Microsoft plans to integrate MAI-Image-1 into its imager as well as Copilot. At the moment, the model is still being tested on the public platform LMArena, “to collect information and feedback.”
Source: BFM TV
