Eleven v3 (Alpha) Released: A New Breakthrough in AI Audio

Eleven Labs has unveiled a major advancement in its AI audio technology — the Eleven v3 (Alpha) model, featuring enhanced emotional expressiveness and multilingual support, designed to deliver an exceptional voice generation experience for creators and developers. This model enables users to flexibly control voice emotion, speed, and effects through audio tags (such as `[laughs]`, `[whispers]`, `[excited]`, etc.), making it ideal for multi-character dialogues and content creation in multiple languages.

Key highlights of the model include support for over 70 languages, richer emotional expression, and adjustable stability settings to accommodate various use cases. Additionally, Eleven v3 recommends leveraging natural language and punctuation to improve output quality, making it suitable for applications such as audiobooks, podcasts, and game character voices.