Eleven v4
Audio modelCard summarizedOur most emotive text-to-speech model yet
● Re-checked 12 Oct 2026 · 48 facts · 0 corrections · 4 unconfirmedSource: Eleven v4 launch post and Eleven v4 docs page (no model card published)
Where to call it
- ElevenLabs API
eleven_v4
Good for
- Character voiceovers for gaming and animation
- Emotional, multi-speaker dialogue where speakers respond to context
- Audiobook production and long-form narration
Overview
ElevenLabs · 02TTS, front
Eleven v4
Our most emotive text-to-speech model yet
Typical price
$0.08 / 1K chars
Eleven v4
- Max length
- ~10 min (10,000 characters)
- Languages
- 90+ languages (87 listed)
- Modes
- text-to-speech, multi-speaker dialogue
- Released
- 28 Sep 2026
Eleven v4 is ElevenLabs' text-to-speech model built on a new architecture, with a low-latency variant, Eleven v4 Turbo. ElevenLabs says it improves on Eleven v3 in output quality, voice accuracy, consistency, emotion, delivery, audio tags and language coverage, and that it follows inline audio tags and direction prompts more accurately than earlier models. launch post; v4 docs
eleven_v4- Typical price
- $0.08 / 1K chars
- Max length
- ~10 min (10,000 characters)
- Languages
- 90+ languages (87 listed)
- Weights
- Closed
- Released
- 28 Sep 2026
- text-to-speech
- multi-speaker dialogue
- audio tags
- voice cloning (instant and professional)
- streaming (v4 Turbo over WebSocket)