GPT-Realtime-2

reasoningproprietary

GPT-Realtime-2 is our most capable realtime voice model. It supports speech-to-speech interactions with configurable reasoning effort, stronger instruction following, and more reliable tool use for complex voice-agent workflows.

Intelligence
Reasoning
Speed
Context
128K
Max output
32K
Input price
$4/1M tokens
Output price
$24/1M tokens

Capabilities

Reasoning & Interaction
Reasoning
Tool Calling

Details

Provider OpenAI
Creator openai
Familygpt-realtime
Licenseproprietary
Statusactive
Typereasoning
Input modalitiestext, audio, image
Output modalitiestext, audio
Open weightNo
Parameters—
Architecture—
Context window128K tokens
Max input—
Max output32K tokens
Reasoning tokensYes
Knowledge cutoff
Release date—
Sourceofficial

Tools

Function Callingfunction_calling
Call external functions and APIs

Endpoints

RealtimeWS
Low-latency streaming via WebSockethttps://api.openai.com/v1/realtime

Pricing

Audio tokensPer 1M tokens
InputCached inputOutput
Standard$32$0.4$64
Text tokensPer 1M tokens
InputCached inputOutput
Standard$4$0.4$24
Image tokensPer 1M tokens
InputCached inputOutput
Standard$5$0.5—

GPT-Realtime-2 supports configurable reasoning effort. Higher reasoning effort can increase latency and output token usage.

Family Comparison: gpt-realtime

ModelMax durationSample rateChannelsLanguagesGenresPricing
GPT-Realtime-Translate——————
GPT-Realtime-Whisper——————
GPT-Realtime-2.1 Mini—————$0.6$2.4
GPT-Realtime Mini—————$0.6$2.4
GPT-Realtime-1.5—————$4$16
GPT-Realtime-2.1—————$4$24
GPT-Realtime-2—————$4$24
GPT-Realtime—————$4$16

API

GET/v1/models/openai/gpt-realtime-2