VibeVoice can produce up to 90 minutes of synthetic dialogue with as many as four distinct speakers.

MAI noted that these models represent the first steps in a larger roadmap. The company plans to orchestrate a range of specialised AI models designed for different use cases and user intents, ...