Eleven v4 at a glance
Eleven v4 targets expressive narration and dialogue, while v4 Turbo is tuned for voice agents with a reported median time to first speech of about 150 ms.
What changed with Eleven v4
ElevenLabs released Eleven v4 on September 28, 2026, calling it its most emotive text-to-speech model and saying it runs on a new architecture. Delivery can be directed with natural-language descriptions and inline audio tags, multi-speaker dialogue reacts more to context, and pronunciation control through IPA phonemes is improved.
Eleven v4 Turbo is the low-latency variant for voice agents and is optimized to work with ElevenAgents. ElevenLabs reports a median inference latency of about 100 ms and a median time to first speech of about 150 ms over WebSocket streaming. Both models support more than 90 languages, instant voice clones from about 10 seconds of audio, and now Professional Voice Clones.
Eleven v4 in numbers
Values are copied from the linked source on the date shown. They are not IA Compa measurements.
- Median time to first speech (v4 Turbo, vendor-reported)as of · Source for Median time to first speech (v4 Turbo, vendor-reported): elevenlabs.io (opens in a new tab)
- ~150 ms
Where to compare Eleven v4
Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.
General leaderboards
Official resources
Why Eleven v4 matters for businesses
The models are available in ElevenAgents, ElevenCreative, and the API. ElevenLabs cites first place on Artificial Analysis's provider voice leaderboard and blind tests in which listeners preferred v4 about 75% of the time; these comparisons are vendor-reported. For agencies building phone agents, the Turbo latency figures are the ones to test against real call flows.
Put it to work: AI Voice Agents
Sources for this story
- Introducing Eleven v4, our most emotive model (opens in a new tab)Official
ElevenLabs · elevenlabs.io
Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.
- #text-to-speech
- #elevenlabs
- #voice-agents
- #voice-cloning
