ElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28. The new text-to-speech generation focuses on more expressive delivery, stronger speaker consistency and improved voice cloning across more than 90 languages. ElevenLabs' release
Eleven v4 is aimed at content creation and long-form audio through the Text to Dialogue API. Eleven v4 Turbo targets interactive use such as voice agents. ElevenLabs reports median inference latency of about 100 milliseconds for Turbo in its September 28 changelog.
Users can describe how a line should be delivered in natural language or add tags for delivery and sound effects. ElevenLabs says v4 follows those directions more accurately and maintains speaker identity across multi-speaker scenes.
Both models were available at launch. ElevenLabs separately reports about 150 milliseconds from request to audible speech for Turbo in its network-adjusted test; that is a different timing boundary from the 100-millisecond inference figure.