Realtime API (public beta)
Realtime API lets developers build low-latency speech-to-speech apps over WebSocket on GPT-4o; audio costs about 6 cents per minute in and 24 cents out.
Streaming audio in and out over one connection, announced at DevDay 2024 with prompt caching (50% off cached input), vision fine-tuning and model distillation tools.
- Date
- Tuesday, 1 October 2024
- Lab
- OpenAI
- Kind
- feature
- Access
- closed API
- Price
- audio $100/$200 per 1M tokens at launch
Figures
| Measure | Value | Measured by |
|---|---|---|
| Audio token price | $100 input / $200 output per 1M tokens about 6 and 24 cents per minute | company |
Sources
This record was checked against its sources on 6 October 2026. How we check