Together AI has released MiniMax Speech 2.8 Turbo, an enterprise text-to-speech model designed for natural-sounding voice agents. This new model offers significant improvements in prosody, includes sound tags for vocal cues like laughter and sighs, and boasts high-fidelity voice cloning capabilities. It also provides end-to-end generation in under 250 milliseconds and is now available on Together AI's dedicated infrastructure, alongside over 600 new voices. AI
Summary written by gemini-2.5-flash-lite from 3 sources. How we write summaries →
IMPACT Enhances the naturalness and expressiveness of AI voice agents, potentially improving user interaction in applications.
RANK_REASON Release of a new speech synthesis model with performance improvements and new features.