Overview
GPT-Realtime-2.1 mini is a distilled reasoning model designed for faster, lower-cost realtime voice interactions. OpenAI announced its release on July 6, 2026. (developers.openai.com)
Capabilities and uses
- Supports audio and text input and output, plus image input. It can be used for realtime voice applications over WebRTC, WebSocket, or SIP.
- Supports function calling and reasoning tokens. OpenAI reports improved alphanumeric recognition compared with GPT-Realtime-2.
- Context window: 128,000 tokens; maximum output: 32,000 tokens. Knowledge cutoff: September 30, 2024. (developers.openai.com)
Limitations
Video, structured outputs, fine-tuning, and predicted outputs are not supported. The model page also lists the generic streaming feature as unsupported; its realtime voice capabilities are described separately. The free API tier is not supported. (developers.openai.com)
API pricing
OpenAI's published prices are in USD per 1 million tokens, checked November 16, 2026. Cached-input prices apply to eligible cached input. (developers.openai.com)
| Modality | Input | Cached input | Output |
|---|---|---|---|
| Text | $0.60 | $0.06 | $2.40 |
| Audio | $10.00 | $0.30 | $20.00 |
| Image | $0.80 | $0.08 | Not supported |
Eligible regional-processing endpoints incur a 10% surcharge under OpenAI's published regional-processing policy. Tool use may incur separate charges. (developers.openai.com)