ElevenLabs announced Eleven v4 and Eleven v4 Turbo for text-to-speech. The company emphasizes emotional nuance, voice cloning, and support for more than 90 languages, with Turbo targeting real-time agents. The models are offered across ElevenLabs products and its API.
Context
Narration and live conversation place different demands on a voice. A prepared performance can prioritize expressive range; an assistant also needs timely responses and clean interruption handling. Compare full interactions in the languages you use. Expressiveness should make meaning clearer, without letting a convincing delivery stand in for a correct answer.
Sources & authors
- Eleven v4: Our most expressive text-to-speech AI model yetElevenLabs · September 28, 2026



