GPT-Realtime-2.1

2026-07-06 · Publisher : OpenAI · Domain : Artificial Intelligence

OpenAI releases GPT-Realtime-2.1, an update to its voice model with reasoning, and GPT-Realtime-2.1 mini, a faster and cheaper distilled version.

What changes

  • Better recognition of alphanumeric characters, handling of silence and noise, and of interruptions.
  • GPT-Realtime-2.1 mini brings reasoning and tool use to the mini range, at the same text price as gpt-realtime-mini.
  • Support for WebRTC, WebSocket and SIP for the mini version.
  • Succeeds GPT-Realtime-2, launched in May.

### Specifications - Context: 128,000 tokens, maximum output 32,000 tokens, text, audio and image input; text and audio output. - Knowledge cutoff: September 30, 2024. Configurable reasoning effort. - Availability: Realtime API (v1/realtime). - Identifiers: gpt-realtime-2.1, gpt-realtime-2.1-mini. ### Pricing | Model | Modality | Input | Cached input | Output | | --- | --- | --- | --- | --- | | gpt-realtime-2.1 | Text | $4.00 | $0.40 | $24.00 | | gpt-realtime-2.1 | Audio | $32.00 | $0.40 | $64.00 | | gpt-realtime-2.1 mini | Text | $0.60 | $0.06 | $2.40 | | gpt-realtime-2.1 mini | Audio | $10.00 | $0.30 | $20.00 | Price per million tokens (USD). ### Benchmarks No numerical score published by the vendor in the announcement. ### Key takeaways A reliability update more than a capability one: it addresses the defects that get in the way in production (codes, numbers, background noise). The mini version makes voice reasoning affordable.

Links

Tags : Audio, OpenAI, Realtime, Voix, Agent vocal