Microsoft reveals two new AI models, indicating direct competition with Openai

Microsoft has announced its new AI models, Mai-Voice-1 and Mai-1 preview, which moves away from the dependence on Openai. Mai-Voice-1 generates audio efficiently, and Mai-1-Preview is in the public testing, which is currently 13th place on LMarena. Microsoft announced two new AI models on Thursday (Bloomberg) Microsoft announced its first internal AI models called Mai-Voice-1 and Mai-1 preview on Thursday. Microsoft has largely relied on Openai’s GPT models to stimulate the AI ​​experience on Copilot, but as the technical giant goes into its own AI models, it is increasingly in competition with the chatgpt manufacturer with a number of other AI businesses. What can Microsoft’s new AI models do? Mai-Voice-1 is a natural speech generation model that, according to the company, can generate a full minute sound under a second and only use a single GPU, which according to the company is one of the most effective speech models. The new model is now offering the Copilot Daily and Podcasts and Feature and Microsoft, and it also starts in Copilot Labs. Meanwhile, Mai-1-Preview is a first foundation model of Microsoft and is now available for public tests on LMARENA, Crowddourced Benchmarking Platform for Great Language Models (LLMS). Mai-1-Preview is currently in the 13th position in the LMARENA rankings, far below the like GPT-5, Gemini-2.5 Pro and even Deepseeek R-1 and Grok-3. Mai-1-Preview was pre-trained and trained by Microsoft at 15,000 Nvidia H100 GPUs, which is a bargain compared to the more than 100,000 employed by competitors such as grok to train its models. Microsoft’s AI road map: In an interaction with semafor, Microsoft Ai chief Mustafa Suleyman said the company took some leather from the Open Source community to stretch the ability of its model with minimal resources. “Increasingly, the arts and crafts of training models are to choose the perfect data and do not waste any of your flops on unnecessary signs that your model has not learned much.” Suleyman said. Suleyman also said that Microsoft AI is already working on the next version of its AI models and is working on the largest data centers in the world equipped with Nvidia’s latest GB-200s chips. “We have big ambitions for where we are going. Not only will we pursue further progress here, but we believe that the orchestration of a series of specialized models that will serve and use different user intent will unlock a tremendous value. There will be many more of this team on both fronts in the near future, ‘Microsoft wrote in a blog post.