As of: 2026-10-25 23:49 CET.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/m/eleven-realtime-v1-max/

# Eleven Realtime v1 Max / Mini

ElevenLabs, Eleven Realtime family. Status: current. Type: speech and audio. Released: 2026-10-20.

## How to call it

| Provider | Model id | Endpoint | Docs |
|---|---|---|---|
| ElevenLabs Realtime API (WebSocket, OpenAI Realtime-compatible) | `eleven_realtime_v1_max` | wss://api.elevenlabs.io/v1/convai/realtime | https://elevenlabs.io/docs/eleven-agents/customization/realtime-models |
| ElevenLabs Realtime API | `eleven_realtime_v1_mini` | wss://api.elevenlabs.io/v1/convai/realtime | https://elevenlabs.io/docs/eleven-agents/customization/realtime-models |
| ElevenAgents |  |  | https://elevenlabs.io/realtime |

## Pricing

- Max input: $1.80
- Max output: $9
- Mini input: $0.20
- Mini output: $6
- Connected audio per minute: $0.08
- Unit as published: USD per 1M tokens (input / output) plus $0.080 per additional minute of connected audio on the Realtime API (burst $0.160/min); same per-minute rate for Max and Mini
- Source: https://elevenlabs.io/pricing/api
- Last checked: 2026-10-25

## Specs

- Input: audio, text
- Output: audio, text
- Context window: not published
- Max output: not published
- Open weights: no
- Licence: proprietary

## Where it stands

- No newer speech and audio model in the registry yet.
- Before it from ElevenLabs: Eleven v4 / Eleven v4 Turbo (2026-09-28) https://postcutoff.com/m/elevenlabs-v4/
- Closed weights; access is hosted only.

## What stands out

- Reasoning speech-to-speech for tool-heavy agents (Max): Max reasons before responding; 87.1% task success on tool-use tasks (Mini 60.3%) and 98.8% on Big Bench Audio (Mini 75.9%), per ElevenLabs citing Artificial Analysis' independent evaluation. Source: https://elevenlabs.io/blog/realtime
- Drop-in OpenAI Realtime API compatibility: The Realtime API follows OpenAI's Realtime spec, so the OpenAI SDK works by changing endpoint and model name; WebSocket only at launch (no WebRTC). Source: https://elevenlabs.io/blog/realtime
- 90+ languages and 17,000+ library voices: Realtime conversation in over 90 languages with any of 17,000+ ElevenLabs voices; time to first audio ~0.57 s (Max) and ~0.62 s (Mini). Source: https://elevenlabs.io/realtime
- #6 on Artificial Analysis Speech-to-Speech Index: Max scored 77.3 on the AA Speech-to-Speech Index (#6 of 21), Arena Elo 997, time to first audio 0.57 s; Mini 64.4 (#17), read 2026-10-25. Source: https://artificialanalysis.ai/speech-to-speech

## Notes

Artificial Analysis speech-to-speech (https://artificialanalysis.ai/speech-to-speech), read 2026-10-25: v1 Max #6 of 21 by Speech-to-Speech Index (77.3), Arena Elo 997, TTFA 0.57 s, $4.80 per hour of input audio (AA-normalised); v1 Mini 64.4 (#17). ElevenLabs' own Speech Agent Arena figures: Max 998, Mini 988; Full-Duplex-Bench Mini 95.8%, Max 94.1%. Mini is the default, unavailable under Zero Retention Mode / HIPAA; Max is available under both (US/EU) but not in India or Singapore data residency, and needs max_tokens >= 1024 or no limit. In ElevenAgents, Realtime mode bills the standard ElevenAgents per-minute rate plus tokens; LLM, TTS model and ASR are managed by ElevenLabs; not usable with custom LLM, Bedrock or Speech Engine. Docs say Realtime models are 'rolling out gradually'. Pricing page does not split audio vs text tokens. No timeline entry yet (lead in data/leads.md, 2026-10-25). Related: elevenlabs-v4, elevenlabs-v3-conversational, gpt-realtime-2-1.

ElevenLabs' first native speech-to-speech models (Oct 20, 2026): Max for complex multi-step agents, Mini for low-cost conversation.

Sources: https://elevenlabs.io/blog/realtime · https://elevenlabs.io/realtime · https://elevenlabs.io/docs/eleven-agents/customization/realtime-models · https://elevenlabs.io/docs/eleven-agents/guides/openai-realtime-compatibility · https://elevenlabs.io/pricing/api · https://artificialanalysis.ai/speech-to-speech

Model page: https://postcutoff.com/m/eleven-realtime-v1-max/