GPT Realtime 2

Realtime voice OpenAI Verified Released May 7, 2026

GPT-Realtime-2 is designed for voice agents that exchange speech with minimal delay. Unlike GPT-Realtime-1.5, which does not reason, it can think through a task before replying aloud or using tools. Summary written by AI Index from OpenAI’s own pages.

  • In Input modalities: TextImageAudio Out Output modalities: TextAudio
  • Reasoning· minimal, low, medium, high, xhigh
  • Tool use

Prices

2 offerings · prices as listed
OpenAI API (opens in a new tab) Official Global · USD gpt-realtime-2
Conditions/ 1M tokensInputCache readOutput
Text tokens Standard $4.00$0.40$24.00
Image tokens $5.00$0.50—
Audio tokens $32.00$0.40$64.00

Prices from OpenAI API pricing (opens in a new tab) (checked Nov 9, 2026)

Standard: the price a request with no conditions pays (the headline price elsewhere on the site).

Details

Max output
32K
Structured output
—
Knowledge cutoff
Sep 2024
Model ID

Not stated on OpenAI’s pages: max input, reasoning can be off, built-in tools, voices, languages.

Cited scores

From third-party leaderboards, not measured by us

No cited leaderboard covers realtime voice models yet. We will cite one when its license allows.

No changes recorded yet.

Sources

OpenAI’s official policies: data use, privacy, terms →

OpenAI · United States · openai.com

Snapshots, the same family and other models from this provider