ElevenLabs published Eleven v4 on September 28, 2026. It describes the model as its most emotive text-to-speech yet, built to interpret tone, pacing, emotion, character and context.
What to know
- Eleven v4 targets expressive narration and character voices.
- Eleven v4 Turbo brings the technology to low-latency use such as voice agents, with a median inference latency of around 100 milliseconds, according to ElevenLabs.
- Both support more than 90 languages, and a voice recorded in one language can speak another while keeping the speaker’s identity.
- Voice cloning is improved, and ElevenLabs says Instant Voice Clones can work from about 10 seconds of audio.
- Both models are available in ElevenAgents, ElevenCreative and the ElevenLabs API.
Only clone a voice with the speaker’s permission and the right to use it.
Source: ElevenLabs blog. We summarise in our own words; check the source for the full details.