ElevenLabs
Eleven v4 is ElevenLabs’ new text-to-speech model for more expressive speech, while Eleven v4 Turbo applies the same research to faster, real-time conversations.
Both models run on a new architecture. They interpret tone, pacing, emotion, character and context, with inline tags giving creators and developers direct control over delivery, reactions, sound effects and style.
The ElevenLabs LinkedIn announcement also lists improvements to voice cloning, long-form speaker consistency, IPA support, multi-speaker dialogue and performance across more than 90 languages. Eleven v4 ranks first in the Artificial Analysis benchmark cited in the post.
Eleven v4 Splits Quality From Real-Time Speed
Eleven v4 is the quality-focused model. ElevenLabs describes it as its most emotive voice model, with more control over how a line is performed rather than simply how clearly each word is spoken.
Eleven v4 Turbo is optimised for live conversations. A second launch post reports roughly 100 ms time to first byte and roughly 150 ms time to first audio. The main announcement gives its median inference latency as approximately 100 ms.
| Model | Main focus | Reported latency | Introductory API price |
|---|---|---|---|
| Eleven v4 | Expressive, high-quality speech | Not stated | $22 per 1M characters |
| Eleven v4 Turbo | Real-time conversations and agents | ~100 ms TTFB, ~150 ms TTFA | $11 per 1M characters |
- Languages
- 90+
- improved performance across supported languages
- Instant clone input
- 10 seconds
- audio needed to capture a voice
- Turbo TTFB
- ~100 ms
- reported time to first byte
- Turbo TTFA
- ~150 ms
- reported time to first audio
Voice Cloning and Delivery Controls Get an Upgrade
Instant Voice Clones can capture a voice from 10 seconds of audio. Professional Voice Clones remain the higher-fidelity option, with the new models intended to make cloned speech more authentic and consistent.
The models also support multi-speaker dialogue. This allows separate characters to appear within the same generated exchange, while the improvements to long-form consistency are designed to keep speakers stable over longer recordings.
Inline tags add more direct control inside the text being generated. They can specify emotion, pacing, delivery, reactions, sound effects and style. IPA support provides phonetic control over pronunciation, including names or specialist terms that a model may otherwise read incorrectly.
These controls apply to content production as well as conversational agents. Eleven v4 is available in ElevenCreative, where it follows the broader production workflow introduced with ElevenLabs Studio 4.0. The models are also available through ElevenAPI and ElevenAgents.
Availability and Introductory Pricing
Eleven v4 and Eleven v4 Turbo are available now in ElevenCreative, ElevenAgents and ElevenAPI. The launch offer runs for two weeks, pricing API generation at $22 per 1 million characters for Eleven v4 and $11 per 1 million characters for Eleven v4 Turbo.
The free allowance described in the announcement applies to Eleven v4, not v4 Turbo.
The agent focus matches ElevenLabs’ wider enterprise use. In a follow-up LinkedIn post about its employee tender offer, the business reported that enterprise customers account for 55% of revenue and are deploying ElevenAgents across sales, support and operations. It also reported use in daily operations at five of the ten largest technology companies, five of the ten largest insurers and four of the ten largest telecoms companies.
Frequently asked questions
What is ElevenLabs Eleven v4?
Eleven v4 is a text-to-speech model built for expressive, controlled speech. It interprets tone, pacing, emotion, character and context, and supports inline controls, voice cloning, multi-speaker dialogue and more than 90 languages.
Is Eleven v4 free?
Eleven v4 is free temporarily in ElevenCreative for Creator+ plans, up to twice the plan’s monthly credits. For two weeks, API access costs $22 per 1 million characters for Eleven v4 and $11 per 1 million characters for v4 Turbo. Standard pricing after the offer was not included in the LinkedIn announcements.
How does Eleven v4 compare with v4 Turbo?
Eleven v4 focuses on maximum speech quality and emotional delivery. Eleven v4 Turbo is optimised for real-time conversations, with roughly 100 ms time to first byte and 150 ms time to first audio.
When is Eleven v4 available?
Eleven v4 and Eleven v4 Turbo are available now. Both can be used in ElevenCreative, ElevenAgents and ElevenAPI.
Sources
6 checkedHow we cover tool news: Create With's tool desk drafts these reports with AI from the sources listed above and checks them against those sources before publishing.
Vercel
PostHog
Figma
Descript
Bubble