Skip to content
Realtime engine · openai-live

The openai-live engine on Wixzel Voice: OpenAI GPT-Live + GPT-6 Luna

OpenAI GPT-Live handles the voice: it listens and speaks at the same time and manages turn-taking. The agent's OpenRouter language model, GPT-6 Luna by default, answers every question and runs every tool, so any OpenRouter model can sit behind the voice. It has 13 built-in voices and follows the caller's language with auto, or speaks one of 32 pinned languages.

Last updated

$0.0827per connected minute, including the $0.012 orchestration fee
Create an accountVoice engines in the docs

What does each model in the openai-live engine cost on Wixzel Voice?

A realtime voice model listens and speaks, and the language model the agent picks writes each answer. Most of the price is the voice model’s seconds. The pipeline label is OpenAI GPT-Live + GPT-6 Luna.

StageModelUnit pricePer minute
Realtime modelopenai/gpt-live-1$0.0012 per second$0.0700
Language modelopenrouter/openai/gpt-6-luna$0.0001 per 1k tokens in$0.0003
Language modelopenrouter/openai/gpt-6-luna$0.0007 per 1k tokens out$0.0003
Orchestrationwixzel/orchestration$0.0120 per minute$0.0120

Rounded up, never down. Where two models are listed for one stage the engine picks by language; the engine price above assumes the dearer path. All prices on the model price list.

Frequently asked questions

What is the openai-live voice engine?

openai-live is a realtime voice engine on Wixzel Voice. OpenAI GPT-Live handles the voice: it listens and speaks at the same time and manages turn-taking. The agent's OpenRouter language model, GPT-6 Luna by default, answers every question and runs every tool, so any OpenRouter model can sit behind the voice. It has 13 built-in voices and follows the caller's language with auto, or speaks one of 32 pinned languages.

How much does the openai-live engine cost per minute?

$0.0827 per connected minute at typical speech rates, including the flat $0.012 per minute orchestration fee. The figure is generated from the price book the meter bills against and is rounded up, never down. Telephony runs over your own SIP trunk and is not included or marked up.

Which models does the openai-live engine use?

openai-live listens and speaks through one realtime model, and a language model the agent picks writes every answer: openai/gpt-live-1, openrouter/openai/gpt-6-luna. There is no separate speech-to-text or text-to-speech stage to configure.

Do I need my own openai account to use openai-live?

No. The provider accounts belong to Wixzel Voice. One wv_live_… API key and a prepaid balance cover the openai-live engine and every other engine; there are no provider keys to bring.

How do I use the openai-live engine in a call?

Set the agent's voice to {"realtime": {"model": "openai/gpt-live-1"}} when you create it with POST /v1/agents, then place a call with POST /v1/calls. GET /v1/engines lists what is available right now; an engine whose provider is degraded disappears from that list before calls start failing.

Is the openai-live price per minute a flat rate?

No. $0.0827 is what a typical minute adds up to. The meter charges per second of speech, per token and per character for the models that actually ran, plus the flat orchestration fee, and every micro-dollar is itemised in GET /v1/usage/events. A quieter or shorter turn costs less; a long monologue costs more.

Which other voice engines does Wixzel Voice run?

classic: $0.0865 per minute

Deepgram + GPT-6 Luna + ElevenLabs

sarvam: $0.0617 per minute

Sarvam, Indian languages end to end

gemini-live: $0.0491 per minute

Gemini Live, native audio

deepgram-agent: $0.1851 per minute

Deepgram Voice Agent

Compose your own

Name a model for each stage and see the minute it adds up to.

Build your first voice agent

Sign upRead the quickstart