Eleven v4 & v4 Turbo by ElevenLabs: status and timeline

Eleven v4 & v4 Turbo

Converts text to natural speech in 90+ languages, clones voices from minimal audio, and controls emotional tone through inline expression tags.

ScalingStage 5 of 5

Released Sep 2026 with 90+ languages and 10-second voice cloning; in API, ElevenCreative and ElevenAgents.

Updated 13d agoChecked 10 Oct0 updates this week

Milestones

No announced next step

ElevenLabs founded2022Complete.
Series D funding at $11B valuation5 Feb 2026Complete.
v4 and v4 Turbo models released28 Sep 2026Complete.
Valuation reaches $22B via share saleJul 2026Complete.

Most important, last 3 months

ElevenLabs releases v4 and v4 Turbo speech models with 90+ language support and 10-second voice cloning13d ago

Current obstacles

Naturalness at scaleAs language coverage grows, maintaining natural-sounding speech across diverse phonetic systems and cultural speech patterns remains challenging.

Physics limits

Information density in audioSpeech conveys less information per second than text or images; this limits how much nuance can be captured or conveyed compared to human-produced content.

How it works

3 parts

Voice cloning

Learn from minimal audio

The model captures speaker identity from just 10 seconds of audio sample, preserving voice characteristics while generating new speech in any language.

Expression control

Stackable inline tags

Creators use inline expression tags to specify emotion, pacing and tone directly in the text, allowing fine-grained control over delivery without re-recording.

Multilingual

Cross-language identity

When synthesizing in languages other than the source voice, the model adopts native accent while maintaining the speaker's unique vocal identity.

Spec sheet

Spec	Eleven v4 & v4 Turbo
Supported languages	90+^R (reported)
Voice cloning minimum sample	10 seconds^R (reported)
Annualized run rate (2026)	$600M+^R (reported)

^R reported by the company

Update log

1 update

Mon 28 Sep

Major: ElevenLabs releases v4 and v4 Turbo speech models with 90+ language support and 10-second voice cloningPressSoftware

About ElevenLabs

The team behind Eleven v4 & v4 Turbo

ElevenLabs

Voice AI, speech synthesis, voice cloning, multilingual

ElevenLabs builds AI speech synthesis models that convert text to natural-sounding speech in over 90 languages and can clone a voice from just 10 seconds of audio. The company serves millions of creators, developers and enterprises through its API, web platform and ElevenAgents.

Raised $500M at $11B in Feb 2026, doubled to $22B mid-2026. Over $600M ARR, 55% enterprise.
v4 and v4 Turbo support 90+ languages with 10-second voice cloning and expression control.
Serves millions via API, ElevenCreative web app, and ElevenAgents voice products.

elevenlabs.io (opens elevenlabs.io)Company profile →

Founded: 20224 yrs
Headquarters: United States
Status: Private
Valuation: $22BprivateJul 2026
Raised: $500M1 round
Last round: Series D · $500M5 Feb 2026
Works in: AIImage & voice generation
Coverage: 1 program · 3 updateslatest 13d agochecked 10 Oct
Lead investors: Sequoia Capital
Investors: Sequoia Capital
People: Mati StaniszewskiCEO & Co-founderPiotr DąbkowskiCTO & Co-founder

Spec

Eleven v4 & v4 Turbo

Supported languages

90+^R (reported)

Voice cloning minimum sample

10 seconds^R (reported)

Annualized run rate (2026)

$600M+^R (reported)