Eleven v3 (alpha) announcement

https://elevenlabs.io/blog/eleven-v3

Expressive TTS with inline audio tags ([whispers], [sighs], [excited]), a Text to Dialogue endpoint for multi-speaker turns and interruptions, 70+ languages. Not recommended for real-time or conversational use at launch (latency, reliability); v2.5 Turbo or Flash recommended for real time.

Realtime speech-to-speech and expressive TTS for characters