Skip to main content

openai/gpt-realtime-2.1/realtime

GPT-Realtime-2.1 reasons over spoken requests, supports function tools and responds with streaming speech or text.

Text · Input $0.80Text · Output $4.80Audio · Input $6.40Audio · Output $12.80Image · Input $1per 1 million tokens80% below the publisher's list priceText · Cached input $0.08Audio · Cached input $0.08Image · Cached input $0.10OpenAI official priceText · Input $4Text · Output $24Audio · Input $32Audio · Output $64Image · Input $5

Workbench

Input

Realtime conversation
34/4096
For a single GPT-Realtime-2.1 task, supply the text that starts one spoken response; live audio is sent through the WebSocket protocol. at least 1 characters · at most 4096 characters
Choose a GPT-Realtime-2.1 preset voice before generating audio; changing a voice after speech begins requires a new conversation.
Reasoning effort
MinimalLowMediumHighExtra high
GPT-Realtime-2.1 accepts minimal, low, medium, high and xhigh reasoning; increased effort can use more output tokens.
Safety checkerOff

SpicyAPI does not run a safety checker on this model. Your prompt and media go to the model as written, and the model itself decides what comes back.

Checking your account before enabling submission and pricing.

Output

Ready for a new run
Example result
0:00 / –:––

Related models and next tasks

Missing the model you need?

GPT-Realtime-2.1 reasons over spoken requests, supports function tools and responds with streaming speech or text.

Key numbers

Context window
128K

Token budget

Input window
128,000
tokens in one request
Longest reply
32,000
25% of the window

reasoning_effort

  • minimal
  • low
  • medium
  • high
  • xhigh

What it does

Reasoned voice turns with function tools

The 2.1 release targets recognition of letters and numbers, handling of pauses and background noise, and behavior when a speaker interrupts. Those capabilities matter when an assistant must confirm a reference code or listen through an unfinished sentence. Recognition still needs application-level confirmation for consequential details; the model’s audio transcript is an aid to the interaction rather than proof that every digit was understood. Configurable reasoning effort lets a voice application allocate more work to a difficult turn. Function tools can connect the conversation to application actions, while streaming audio lets playback begin before a whole answer is ready. The model has a 128,000-token context window and up to 32,000 output tokens; each connection uses the smaller context and output budgets selected by the caller.

Where it fits

  • Scheduling by voice

    For a spoken scheduling assistant using GPT-Realtime-2.1, let a function check availability before the assistant proposes a time. Read back the selected date and participants before a separate application action confirms the booking.

  • Guided device troubleshooting

    For guided troubleshooting, keep the current device state in the conversation and ask one diagnostic question at a time. Use a tool result to establish whether an operation succeeded, then let the model explain the next step aloud.

Made with GPT-Realtime-2.1

  • Spoken confirmation
  • Ferry departure reminder
  • Restaurant booking confirmation
  • Noon meeting reminder

Model reference

GPT-Realtime-2.1 API reference

How to call the GPT-Realtime-2.1 API

  1. Prompt guide & Pro tips

    Give GPT-Realtime-2.1 a concise role, the desired language and a rule for confirming ambiguous names or numbers. Describe when each function should be called and which result permits a success statement. Start with minimal reasoning for a short acknowledgement, then compare a higher effort on representative multi-step questions instead of assuming every turn benefits from extra reasoning. Choose a voice, playback encoding and speaking speed before a response starts. PCM is 24 kHz mono audio; G.711 uses 8 kHz. With automatic voice detection, tune the silence threshold against realistic microphone noise and short hesitations. The client controls playback and interruptions. This endpoint uses server WebSocket transport; its audio input transcription runs on a separately metered transcription connection. Reserve enough history for useful context while keeping the per-turn output cap appropriate to a spoken answer.

GPT-Realtime-2.1 API questions

All GPT models18