MiniMax Audio launches Speech-02: supports 30+ languages, 200,000 characters input

MiniMax Audio officially released the Speech-02 series voice models, supporting 30+ languages ​​and breaking through the single input of 200,000 characters, further improving the naturalness, fluency and practicality of speech synthesis.

The core highlights of Speech-02 include:

• Up to 99% of the voice similarity, the synthesized voice is more natural and realistic;

• Zero rhythm failure completely eliminates the problems of voice lag and rhythm instability;

• "Read Anything" function supports one-click conversion of text, web page URL, and e-books to voice;

• "Long-Text Mode" mode, processing ultra-long text at one time, is suitable for creating content such as audio books, podcasts, etc.;

• Enhanced historical management functions, supporting convenient viewing and organizing voice synthesis records;

• The new "Discovery Hub" helps users quickly master new features of the platform.

MiniMax Audio lowers the threshold for high-quality voice synthesis through Speech-02, allowing users to create and obtain audio content more easily.

At present, the model has been launched, and users can go to the MiniMax Audio official website to experience it.

Experience entrance: https://www.minimax.io/audio