The gemini-live engine on Wixzel Voice: Gemini Live, native audio
One model for the whole turn. Google Gemini Live takes the caller's audio in and produces audio out with the reasoning inside the provider, so there is no separate speech-to-text or text-to-speech stage to name and latency is what the model itself delivers.
Last updated
What does each model in the gemini-live engine cost on Wixzel Voice?
A realtime engine hands the whole turn to one model, so the price is mostly that model’s units. The pipeline label is Gemini Live, native audio.
| Stage | Model | Unit price | Per minute |
|---|---|---|---|
| Realtime model | google/gemini-live | $0.0042 per 1k audio tokens in | $0.0063 |
| Realtime model | google/gemini-live | $0.0168 per 1k audio tokens out | $0.0252 |
| Realtime model | google/gemini-live | $0.0007 per 1k tokens in | $0.0017 |
| Realtime model | google/gemini-live | $0.0028 per 1k tokens out | $0.0013 |
| Orchestration | wixzel/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 gemini-live voice engine?
- gemini-live is a realtime voice engine on Wixzel Voice. One model for the whole turn. Google Gemini Live takes the caller's audio in and produces audio out with the reasoning inside the provider, so there is no separate speech-to-text or text-to-speech stage to name and latency is what the model itself delivers.
- How much does the gemini-live engine cost per minute?
- $0.0466 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 gemini-live engine use?
- gemini-live hands the whole turn to one realtime model:
google/gemini-live. There is no separate speech-to-text or text-to-speech stage to configure. - Do I need my own google account to use gemini-live?
- No. The provider accounts belong to Wixzel Voice. One
wv_live_…API key and a prepaid balance cover the gemini-live engine and every other engine; there are no provider keys to bring. - How do I use the gemini-live engine in a call?
- Set the agent's voice to
{"realtime": {"model": "google/gemini-live"}}when you create it withPOST /v1/agents, then place a call withPOST /v1/calls.GET /v1/engineslists what is available right now; an engine whose provider is degraded disappears from that list before calls start failing. - Is the gemini-live price per minute a flat rate?
- No. $0.0466 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.
What do developers build with the gemini-live engine?
Support calls are conversational and interruption-heavy; a realtime engine keeps the turn latency lowest. The classic engine is the cheaper choice if cost matters more than snappiness.
One realtime model handles listening, thinking and speaking, so the round trip is what Gemini itself delivers and it is the cheapest engine per minute. The classic and sarvam pipelines run over the socket too; choose them when you need a specific voice or an Indian language.
Which guides build on the gemini-live engine?
Let users talk to your voice agent inside a Flutter app: your server mints a session, the `wixzel_voice` Dart SDK streams audio over a WebSocket, and there is no SIP trunk or number.
Put Google's Gemini Live model on a real phone call over your own SIP trunk with four cURL requests, then read the call record with its transcript and cost.