You cannot find Live
Update ChatGPT, verify region, plan and workspace settings, then check Settings → Voice for Live, Advanced, or Standard.
Direct answers
GPT-Live FAQ
GPT-Live is OpenAI's full-duplex ChatGPT Voice model family. Use the answers below for current setup, plans, API status, video limits, mode differences, and troubleshooting.
Last verified:
Start here
GPT-Live is available in supported ChatGPT Voice experiences. Free uses GPT-Live-1 mini, paid consumer plans use GPT-Live-1, video remains in Advanced Voice, and the GPT-Live-1 developer API is not generally available.
See the July updateFive questions in one minute
| Question | Short answer | Best next guide |
|---|---|---|
| What is GPT-Live? | A full-duplex voice model family inside ChatGPT | What is GPT-Live |
| How do I use it? | Open Voice on a supported ChatGPT surface | How to use GPT-Live |
| Is it free? | Yes, with limited GPT-Live-1 mini access | Is GPT-Live free |
| Does it have an API? | No generally available GPT-Live-1 API yet | API release status |
| Does it support video? | Not in Live at launch; use Advanced Voice | GPT-Live video |
Common practical scenarios
You need a camera
Switch to Advanced Voice on an eligible supported mobile account; visual cards and still images in Live are not a live camera stream.
You are building a voice app
Use the current Realtime API rather than copying ChatGPT behavior or guessing a GPT-Live model ID.
GPT-Live questions and answers
What is GPT-Live?
GPT-Live is OpenAI's full-duplex voice model family for more natural ChatGPT Voice conversations. It can listen while speaking, handle interruptions, and delegate harder work to frontier models.
How do I use GPT-Live?
Open ChatGPT on a supported web, iOS, Android, or desktop surface, select the Voice control, allow microphone access, choose Live when offered, and begin speaking.
Is GPT-Live free?
Yes. Free users have limited GPT-Live-1 mini access during a rolling 24-hour period. Paid consumer plans add GPT-Live-1 or higher allowances.
Does GPT-Live have an API?
GPT-Live-1 is not generally available through a documented API. OpenAI provides a notification signup, while developers can use GPT-Realtime-2.1 now.
GPT-Live vs Advanced Voice Mode?
Choose Live for full-duplex conversation, interruption handling, and current intelligence features. Choose Advanced Voice when you need supported mobile video or screen sharing.
Does GPT-Live support video or screen sharing?
Not at launch. Eligible subscribers can use camera video and screen sharing in Advanced Voice on supported iOS and Android apps.
How long can one GPT-Live conversation last?
One Live conversation can last up to two hours, while the total rolling 24-hour allowance depends on plan.
Why is GPT-Live missing?
Common causes are staged rollout, an unsupported region or app version, workspace settings, the selected Voice mode, or an ineligible surface.
When an answer differs from your account
Treat the account UI and current OpenAI Help Center as more authoritative than a dated article. Availability and limits can change, and workspace administrators can restrict Voice or early model access.
Check the verification date on this page, follow the official sources below, and use the dedicated troubleshooting guide when the symptom is device-specific.