Microsoft AI recently announced the launch of its first text-to-image generator designed and developed in-house - MAI-Image-1. As "the next step in its journey," the release of MAI-Image-1 follows the launch of Microsoft's first internal AI models.
Microsoft stated that MAI-Image-1 incorporated feedback from creative professionals during development, aiming to avoid "repetitive or stylized outputs." The company claims that this new model excels at generating photo-realistic images, especially when handling subjects such as lightning and landscapes. More notably, MAI-Image-1 can process requests and generate images faster than "larger, slower models."
Its performance has received authoritative recognition: MAI-Image-1 ranks in the top ten on the artificial intelligence benchmarking website LMArena, which allows users to compare and vote for the best outputs from different systems.
MAI-Image-1 enriches Microsoft's AI product line, joining previous offerings such as the voice generator MAI-Voice-1AI and the chatbot MAI-1-preview.
As its relationship with OpenAI becomes increasingly complex, Microsoft is heavily investing in training its own AI models, including MAI-Image-1. The company has also started using Anthropic's AI models in certain features of Microsoft 365.
Microsoft emphasized that it is committed to "ensuring safe and responsible outcomes" and has invested "significant funds" in the development of MAI-Image-1. (Note: At the time of writing this article, the model had not yet been tested, and follow-up reports on its safety measures will be provided promptly.)
