ElevenLabs launches Eleven Realtime speech-to-speech models for voice agents
Overview
ElevenLabs launches Eleven Realtime, two speech-to-speech models that each act as a complete voice agent in more than 90 languages.
The Mini model is for natural conversations at low cost, and the Max model is for complex tasks. Both can use any of the 17,000+ voices in the ElevenLabs library, and developers can use them in ElevenAgents or through a WebSocket API compatible with OpenAI's Realtime API.
The company says the models start talking within half a second and answer spoken reasoning questions correctly. Artificial Analysis reports 98.8% accuracy for Max and 75.9% for Mini on Big Bench Audio, and 87.1% tool-use task completion for Max versus 60.3% for Mini. These figures come from a third-party report, while the speed and reasoning claims are ElevenLabs' own.
Written by AI from the articles below · updated Oct 20, 10:53 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
ElevenLabs@ElevenLabsOfficialElevenLabs launches Eleven Realtime speech-to-speech models for voice agentsAIElevenLabs introduces Eleven Realtime, a speech-to-speech model class that it says starts talking in half a second. The company says the models answer spoken reasoning questions correctly, so voice agents need not trade response speed for accuracy.

- ElevenLabs BlogOfficialPickElevenLabs launches Eleven Realtime, two speech-to-speech models for voice agents
AIElevenLabs launches Eleven Realtime v1 Mini and v1 Max, speech-to-speech models that each act as a complete voice agent in more than 90 languages. Artificial Analysis reports 98.8% accuracy for v1 Max and 75.9% for v1 Mini on Big Bench Audio, with v1 Max completing 87.1% of tool-use tasks versus 60.3% for v1 Mini. Developers can use it in ElevenAgents or through a WebSocket API compatible with OpenAI's Realtime API.
Heat Trend
Not enough continuous observations to show a trend yet.