Feature

Playground

Chat, talk and create with any model in your browser before writing code.

The Playground is a full assistant running on your own key: private chat threads, voice conversations with GPT Live, image, video and music creation tools, file inputs and personal memory. Open any model page and press Test in Playground to try that exact model — what you use is what you pay, per request, from your wallet.

Chat
Private threads
Voice
GPT Live
Create
Image · Video · Music

What you get

  • Temporary chats that save nothing
  • Switch models mid-conversation
  • Voice calls with natural interruptions
  • Image, video and music tools built in
  • Billed per request from your wallet

Common use cases

Model evaluationPrompt testingDemosEveryday assistantCreative work

How it works

  1. 1
    Sign in
    The Playground uses your own account, keys and wallet.
  2. 2
    Pick a model
    Choose any enabled model — or open a model page and press Test in Playground.
  3. 3
    Chat, talk or create
    Type, speak, or switch to the image, video and music tools.
  4. 4
    Pay per request
    Each request is metered at the model's live market price.

Frequently asked questions

Is the Playground free?

Requests are billed from your wallet at the same market price as the API — no markup for using the browser.

Are my chats private?

Yes. Threads are private to your account, and temporary chats save nothing at all.

Can I test a specific model?

Yes — every model page has a Test in Playground button that opens that exact model.

Related features

GPT Live voice

Real-time spoken conversations with GPT Live 1 — low latency, natural interruptions.

Every frontier model, one endpoint

GPT-6 Astra, Claude Opus 5.5, Claude Fable 5.1, Gemini 3.8 and more — one key, one wallet.

Transparent billing

Every request shows tokens, cost, latency and model — at market price.