ElevenLabs Realtime

Realtime voice agent models with 570ms to first audio

Visit WebsiteShare

Eleven Realtime is ElevenLabs' new family of speech-to-speech models built for voice agents. Speech recognition, turn-taking, the LLM, and the voice are all tuned end-to-end. The v1 Mini model is built for natural conversation, while v1 Max handles complex tasks. It reaches first audio in 620ms on Mini and 570ms on Max. It supports more than 90 languages and works with ElevenAgents or through an OpenAI-compatible Realtime API, so you only need to change the endpoint and model name.