MiniMax Audio Launches Speech- 02 Speech Model, Capable of Inputting 200,000 Characters at a Time

MiniMax Audio, the much-anticipated audio technology innovator, has officially released its new Speech-02 series of speech models, which support more than 30 kinds of speech and can input 200,000 characters at one time. It brings users a more realistic, smooth and convenient audio experience. What's more surprising is that Speech-02's human voice similarity is as high as 99%, which means that the synthesized voice sounds more natural and close to the real person. In addition, the model also realizes zero rhythmic faults, completely solving the problems of lagging and unstable rhythm that may occur during audio playback, ensuring a coherent and smooth listening experience. It is worth emphasizing that despite the major upgrades in various aspects, the Speech-02 series still maintains its original affordable price. To address the need for long text processing, MiniMax Audio has introduced the powerful "Long-Text Mode", which supports asynchronous speech synthesis of up to 200,000 characters in a single input, making it easier than ever to create audio books, podcasts, and other long audio content. This makes it easier than ever to create long audio content such as audio books, podcasts, etc. It completely solves the problem of segmentation when synthesizing long text in the past.