What happened
In its October 19, 2026 changelog, ElevenLabs introduced Eleven Realtime, a new family of speech-to-speech models for voice agents in ElevenAgents. There are two: Eleven Realtime v1 Mini, the default and cheaper option for natural conversation, and Eleven Realtime v1 Max, which reasons before it answers, for complex instructions and multi-step tool calls, and costs more.
Instead of choosing an LLM, speech-to-text and text-to-speech model separately, you set the agent's Voice mode to Realtime and pick one model; ElevenLabs manages the rest. The system prompt, voice, tools and knowledge base work as before. This is a separate launch from Architect, the assistant for improving voice agents that ElevenLabs released in alpha earlier.
Realtime API and the same-day updates
Eleven Realtime is also available through a Realtime API that, according to ElevenLabs, uses the same events and WebSocket transport as OpenAI's Realtime API, so an existing client needs a new endpoint and session configuration rather than a rewrite. Your backend creates a short-lived client secret and the browser connects directly.
The same changelog added Realtime Translation, a WebSocket API that turns speech in one language into speech in another while the speaker is still talking, available for enterprise workspaces. ElevenLabs' product page reports its own figures, such as 620 ms time to first audio for Mini and 570 ms for Max. That differs from the documentation, which describes Mini as the lowest-latency model, so test both with your own agent; these are vendor results, not independent tests.
Pricing and limits
Realtime models are rolling out gradually; if the Voice mode setting is missing on an agent, your workspace does not have them yet. A Realtime agent is billed at the standard ElevenAgents per-minute rate plus input, cached input and output tokens at the Realtime model's rates, which are listed on the ElevenLabs pricing page.
Realtime mode does not work with a custom LLM or Amazon Bedrock, with text-only agents or with transfers to agents in Custom mode. Mini is not available under Zero Retention Mode or HIPAA, and Max is not available in the India and Singapore data residency environments.
Go to the source
Sources & further reading
Based on the official documentation below. Features and access can change; check the provider before making a decision.
Was this page useful?
Help us make our coverage more useful. This is feedback on the page, not a rating of the tool.
Go a little further