OpenAI · voice model

GPT-Realtime-2.1

Release Ruler for GPT-Realtime-2.1: released Jul 6, 2026, $4 per 1M input tokens, proprietary. Each thin stem on the ruler is another model we track.

GPT-Realtime-2.1 is an OpenAI voice model released Jul 6, 2026; OpenAI's newest voice model in our dataset is GPT-Live 1. Proprietary, API access only. OpenAI's realtime speech-to-speech reasoning model, released Jul 6, 2026 with better alphanumeric recognition and interruption handling. Verified against vendor sources on Oct 26, 2026.

Reported Last verified · All models · JSON · Markdown

  • ReleasedJul 6, 2026
  • Context window128K128,000 tokens
  • Max output32K32,000 tokens
  • LicenceProprietaryProprietary (API)
InTextAudioImageOutTextAudio

Key facts

GPT-Realtime-2.1: key facts with sources
FactValueSource
VendorOpenAIn/a
ReleasedJul 6, 2026source
Statusgenerally availablen/a
API idgpt-realtime-2.1source
Context window128,000 tokens (128K)source
Max output32,000 tokenssource
Inputtext, audio, imagesource
Outputtext, audiosource
LicenceProprietary (API)source
Pricing$4 in / $24 out per 1M tokenssource

Where GPT-Realtime-2.1 sits in our data

How GPT-Realtime-2.1 differs from GPT-Realtime-2.1 mini

Read from the recorded facts of the two models, not from a head-to-head test. We have not run either.

  • Listed price is $4 in and $24 out, against $0.6 in and $2.4 out for GPT-Realtime-2.1 mini (per 1M tokens).
  • Both list a 128K-token context window.
  • Both accept text, audio, image as input.
  • Both were released on Jul 6, 2026.
  • Facts for the other model: GPT-Realtime-2.1 mini.

What we tested

Nothing yet. We have not run GPT-Realtime-2.1 ourselves, so every fact on this page is Reported from the sources below. When we do test it, the scope will be stated here and in our test logs. See methodology.

Sources

More from OpenAI

All 88 models