ElevenLabs Launches v4 Voice Models
ElevenLabs releases new voice models with greater expressiveness, lower latency and support for more than 90 languages.
ElevenLabs is pushing voice AI further into real-time, multilingual and enterprise-ready applications.
What happened
The company launched ElevenLabs v4 and v4 Turbo.
The new models add more expressive control, lower latency and support for more than 90 languages, up from 70 previously. The architecture can clone a voice from roughly 10 seconds of audio and improves contextual expression over longer passages.
ElevenLabs says more than 55% of its business now comes from large enterprises.
Why it matters
Voice AI is moving from simple text-to-speech toward live conversational infrastructure.
For customer service, media, education and accessibility, the important factors are not only naturalness but also latency, emotional control, multilingual quality and reliability.
The enterprise revenue mix suggests the market is becoming less experimental and more operational.
The bigger picture
Voice could become a major interface for AI agents.
As companies deploy agents that speak to customers, employees or audiences, high-quality speech generation becomes part of the core product experience. That also raises new questions around consent, identity and synthetic-voice misuse.