Eleven v4 Turbo

Audio modelCard summarized

Our most expressive, real-time speech synthesis model

Source: Eleven v4 launch post, models page and Text to Dialogue WebSocket guide (no model card published)

Use it via API

$0.04 / 1K chars

Code example Compare

Where to call it

ElevenLabs APIeleven_v4_turbo

Supports

No feature list in the maker's docs.

Full feature support

Good for

Support agents and AI assistants
Interactive characters
Real-time voice agents in ElevenAgents, e.g. healthcare agents saying medical terms, gaming agents using slang

Overview

ElevenLabs · 03TTS

v4 Turbo

Our most expressive, real-time speech synthesis model

Typical price

$0.04 / 1K chars

Eleven v4 Turbo

Max length: —
Languages: 90+ languages
Modes: text-to-speech, streaming (Text to Dialogue WebSocket)
Released: 28 Sep 2026

Eleven v4 Turbo is the low-latency variant of Eleven v4, ElevenLabs' text-to-speech model built on a new architecture. It is aimed at real-time use such as voice agents and interactive characters, and is served over the Text to Dialogue WebSocket and in ElevenAgents, where it became the default voice model. ElevenLabs gives a median inference latency of about 100 ms and a median time to first speech of about 150 ms. launch post; models page; changelog 28 Sep and 5 Oct 2026

eleven_v4_turbo

Typical price: $0.04 / 1K chars
Languages: 90+ languages (same list as Eleven v4)
Weights: Closed
Released: 28 Sep 2026

text-to-speech
streaming (Text to Dialogue WebSocket)
audio tags
voice cloning (instant and professional)
voice agents (ElevenAgents)