Eleven v3 at a glance
Eleven v3 added inline audio tags such as [whispers] and [excited], multi-speaker dialogue and support for 70+ languages, aimed at media rather than real-time agents.
What changed with Eleven v3
ElevenLabs published Eleven v3 (alpha) on June 3, 2025, calling it its most expressive text-to-speech model. Writers can steer delivery with inline audio tags in square brackets, such as whispers, sighs or excitement, and the model covers more than 70 languages.
Alongside the existing text-to-speech endpoint, ElevenLabs added a Text to Dialogue API that turns a structured list of speaker turns into one natural conversation with overlaps and interruptions. The company noted that v3 needs more prompt work than earlier models and recommended staying on v2.5 Turbo or Flash for real-time and conversational use while a real-time v3 was developed. The page now notes that v3 has since left alpha.
Why Eleven v3 matters for businesses
For businesses, v3 raised the bar for narrated videos, audiobooks, ads and training content in many languages, including Spanish. Phone agents still needed lower-latency models, and cloning a real person's voice requires that person's consent.
Put it to work: AI Voice Agents
Sources for this story
- Eleven v3: Most Expressive AI TTS Model Launched (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-ai
- #multilingual
