ElevenLabs Launches New Voice Model v4!
ElevenLabs v4 and v4 Turbo offers more than 90 languages, voice cloning with a 10 second clip, and zero latency between question and answer.
ElevenLabs released a new voice model, v4, and a lighter version called v4 Turbo, on September 28. It now supports over 90 languages instead of 70 in the previous version, with the biggest quality jump in Japanese, Brazilian Portuguese, Mandarin and Cantonese. You can now clone a voice from a 10-second recording instead of the longer sample needed before.
For me the lower latency (delay before the voice starts talking) matters more than the language count: in voice agents, people feel the pause before the answer in the very first second of a conversation. With v4, the audio starts generating as soon as the language model produces the first words of its answer, instead of waiting for the whole text to be ready — this cuts the delay, though it doesn't disappear entirely. Cloning a voice from a 10-second sample instead of a long session is also practical: with a short sample it's easier to quickly set up a voice for a support agent without making someone record a long text.
Source: techcrunch.com
Free course
Stop reading about AI — start building with it
The free Claude Code course: your first site, tool or game — no coding. No upsells, no cross-sells — nothing to buy here.
Start free →
Author
Evgenii Arsentev
PhD · Founder & CEO, Dad4Kids
Articles · Latest articles