Microsoft has launched three new artificial intelligence models targeting some of the most demanding tasks in enterprise and creative workflows: transcription, voice synthesis, and image generation. The announcement signals the company’s push to challenge rivals across every major modality of AI output.
According to Microsoft, the three models, named MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2, are immediately available through Microsoft Foundry, the company’s platform for deploying AI capabilities at scale. All three are competitively priced and designed to outperform existing tools on speed, quality, and benchmark rankings.
MAI-Transcribe-1 is Microsoft’s new speech-to-text model, supporting the top 25 most-used languages globally. According to Microsoft, it delivers 2.5 times faster batch transcription than existing Azure offerings and ranks first on the FLEURS benchmark for 11 core languages. Priced at $0.36 per hour, it is built for organizations running large-scale transcription workloads across multiple languages.
MAI-Voice-1 tackles text-to-speech with a focus on natural, emotionally nuanced speech output. The model can generate 60 seconds of audio in just one second, a significant speed improvement over comparable tools. It also introduces a notable new capability: custom voice creation from just a few seconds of audio input. This positions MAI-Voice-1 as a strong contender for applications in content creation, accessibility tools, and interactive AI assistants.
MAI-Image-2 rounds out the trio with faster, higher-quality image generation. According to Microsoft, the model generates images twice as fast as its predecessor and ranks among the top three on the Arena.ai leaderboard. It has been specifically optimized for accurate skin tones, natural lighting, and clear text rendering, areas where many AI image generators have historically underperformed. Rob Reilly, WPP Global Chief Creative Officer, described MAI-Image-2 as “a genuine game-changer” that “deeply respects the craft involved in generating campaign-ready images.”
The launch comes as competition in AI model development intensifies. OpenAI, Google, and Anthropic have all released major model updates in recent months, with the race now extending beyond large language models into specialized tools for audio, vision, and multimodal tasks.
Microsoft’s strategy with the MAI series appears aimed at building a full-stack AI ecosystem within its Foundry platform. Rather than relying on third-party providers for each modality, the company is building its own end-to-end pipeline, giving enterprise customers access to transcription, voice, and image generation under one roof.
For businesses looking to integrate AI into their operations, the MAI model family offers a compelling combination of performance and cost efficiency. Immediate availability through Foundry, with no waitlists or preview restrictions, is a deliberate move to capture enterprise adoption at speed. The pricing is transparent, the benchmarks are specific, and the use cases are clear, suggesting Microsoft intends these models to be production-ready from day one.
