Eleven v4 Turbo

Eleven v4 Turbo is the real-time variant of ElevenLabs’ Eleven v4 text-to-speech model, announced on 28 September 2026. ElevenLabs says it brings v4’s expressive delivery to latency-sensitive uses such as phone and chat agents, with a median inference latency of about 100 milliseconds and a median time to first speech of about 150 milliseconds; these are the company’s own figures. It is built to run with ElevenAgents, ElevenLabs’ conversational-agent platform, and is also available through the API. Like v4, it supports 90+ languages and inline direction tags, and it is a hosted model with no published weights; the pages we read give no per-character price. For buyers building voice agents the practical test is whether the latency holds on their own telephony route and languages, which a vendor median cannot show.

Not yet independently verified: Every detail here comes from ElevenLabs’ own announcement and model page; no independent report or measurement of Eleven v4 was found. The latency figures and the listener-preference result are ElevenLabs’ own, and its pages give no per-character API price. We will update this when it can be confirmed, and remove this note.