MiniMax Audio just leveled up with the new Speech-02 model! Get ultra-realistic Al voices (30+ langs, 99% similarity). Read Files/URLs & handle long text (200k chars). API available at: [email protected].
Hi everyone!
MiniMax Audio just leveled up with their new Speech-02 voice models! They're pushing for ultra-realistic text-to-speech that sounds incredibly natural and expressive across many languages.
What's new with Speech-02 & the platform:
🗣️ Realistic & Expressive: Aims for "native flair" in 30+ languages, with studio-grade clarity and zero rhythm glitches. Think flawless multi-language transitions and even deep, cinematic bass tones.
📎 Read Anything: Instantly turn any file or URL into lifelike audio.
📜 Long-Text Mode: Effortlessly create audiobooks/podcasts with up to 200,000 characters in a single go.
🎤 Unlimited Voice Cloning: Still offers unlimited voice cloning to personalize your audio.
⚡ Sub-Second Streaming: Delivers audio quickly for real-time applications.
💰 Same Price: All these improvements apparently come at the same affordable price point.
This combination of high-quality, expressive multilingual voices and practical features like file reading and long-text support are really powerful for creators.
MiniMax Audio just leveled up with the new Speech-02 model
🚀 ALL New Speech-02 Series Enhanced authentic multi-language coverage, 99% vocal similarity, zero glitches in rhythm, and studio-grade clarity - all at the same affordable price.
📎 Read Anything Upload files or URLs and enjoy listening to any content in your preferred voices, anytime, anywhere.
📜 Long-Text Mode Effortlessly create audiobooks or podcasts without truncation; supports up to 200,000 characters of asynchronous speech synthesis in a single input.
📂 Enhanced History Management Review, delete, or organize your speech synthesis history and settings with ease.
🔍 New Discovery Hub Discover all features and learn what's new in one convenient place.
MiniMax Audio just leveled up with the new Speech-02 model
🚀 ALL New Speech-02 Series Enhanced authentic multi-language coverage, 99% vocal similarity, zero glitches in rhythm, and studio-grade clarity - all at the same affordable price.
📎 Read Anything Upload files or URLs and enjoy listening to any content in your preferred voices, anytime, anywhere.
📜 Long-Text Mode Effortlessly create audiobooks or podcasts without truncation; supports up to 200,000 characters of asynchronous speech synthesis in a single input.
📂 Enhanced History Management Review, delete, or organize your speech synthesis history and settings with ease.
🔍 New Discovery Hub Discover all features and learn what's new in one convenient place.
The progress in natural and expressive speech with Speech-02 is seriously impressive. MiniMax Audio is setting a new bar for multilingual voice quality.
These are pretty good. I think on par with 11labs. But as with 11labs, very skewed towards audio book narration. I guess thats where the demand is at the moment. I wonder if its possible to create voices that are suitable for productvideos, commercials etc. Or even reviews. Similar to openAI's Alloy voice.
