Eleven v4 and Eleven v4 Turbo are ElevenLabs’ fastest and most emotive voice models yet.

Built on an entirely new architecture, Eleven v4 interprets tone, pacing, emotion, character, and context to generate speech that feels performed. Eleven v4 Turbo brings the same research to low-latency use cases such as agents, generating speech with a median inference latency of ~100 ms.

Voice cloning is more authentic and consistent. Instant Voice Clones capture voices with extraordinary fidelity from just 10 seconds of audio, and Professional Voice Clones offer our highest fidelity. For long-form creation, Eleven v4 preserves speaker identity reliably across infinite generations, perfect for long form narration, dialogue, and regenerated lines.

Direct the performance with inline tags. Specify delivery, emotion, pacing, reactions, SFX, or style. Eleven v4 stages the scene and performs it with naturalness and precision.

What's new:

A completely new Text to Speech architecture built for expressive performance
Eleven v4 Turbo for agents and anything live, at a 100ms median latency
Instant Voice Clones from 10 seconds of audio, Professional Voice Clones for highest fidelity
Improved performance across 90+ languages, with new languages including Cantonese, Mongolian, and Odia
Speaker identity preserved across long-form narration, dialogue, and regenerated lines
More natural multi-speaker dialogue
Pronunciation control with IPA support
Inline tags for direct control over delivery, emotion, pacing, reactions, SFX, and style

Available now in ElevenCreative, ElevenAgents, and ElevenAPI.

Try Eleven v4 → https://elevenlabs.io/v4

P.S. I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified → @rohanrecommends