How to Use GPT-Live in ChatGPT: Tiers, Variants, and CarPlay

GPT-Live is rolling out globally as the new engine behind ChatGPT Voice on iOS, Android, and chatgpt.com. There is no separate app or manual toggle for most users: once the rollout reaches your account, GPT-Live becomes the voice model behind ChatGPT Voice.

Try Apidog today

This guide consolidates the implementation details developers and power users usually have to piece together: which model each plan gets, when to use Medium or High reasoning variants, how CarPlay works, and what to check if GPT-Live is not visible yet.

Which GPT-Live model you get

Your ChatGPT plan determines the default voice model:

Plan	Default voice model	Backend
Free	GPT-Live-1 mini	GPT-5.5 Instant
Go / Plus / Pro	GPT-Live-1	GPT-5.5 Instant
Reasoning variants	GPT-Live-1 Medium / High	GPT-5.5 Thinking, medium / high effort

All variants use a full-duplex interaction model:

• The model can listen while speaking.
• You can interrupt it without restarting the conversation.
• It can backchannel naturally during speech.
• It keeps conversational context across interruptions.

Use the default GPT-Live model for low-latency voice interactions. Use Medium or High when reasoning quality matters more than response speed. These variants run GPT-5.5 Thinking behind the conversation, so they are better suited for deeper work sessions.

Start a GPT-Live conversation

To start using GPT-Live:

1. Update the ChatGPT app on iOS or Android, or open chatgpt.com.
2. Tap the voice icon in the message bar.
3. Start speaking.

There is no push-to-talk flow. GPT-Live listens continuously during the session, and you can interrupt it at any point.

Useful interaction patterns include:

• Ask for live translation between languages.
• Ask about weather, stocks, or sports to trigger visual cards.
• Upload an image or file during the conversation and continue discussing it.
• Use memory the same way you would in text chats.

Not available at launch: video and screen sharing. OpenAI says these capabilities are coming later. If you need them now, Advanced Voice Mode remains the better option. For a detailed comparison, see GPT-Live vs Advanced Voice Mode.

If you do not see GPT-Live yet

The rollout is staged. If your voice mode still behaves like the previous version, check the following:

• Update the ChatGPT app. Older app versions may keep using older voice stacks.
• Check the model indicator in the voice interface. Rollout timing depends on your account, not only your app version.
• Expect Free tier availability later. Free accounts typically receive new OpenAI features after paid accounts.
• Use existing voice modes meanwhile. Standard Voice Mode and Advanced Voice Mode remain available.

There is no waitlist and no manual setting to enable. Once your account is included in the rollout, GPT-Live should appear automatically.

Use GPT-Live in your car with CarPlay

ChatGPT in CarPlay is available globally on all plans. It pairs well with GPT-Live because full-duplex voice reduces turn-taking friction while driving.

Requirements:

1. An iPhone running iOS 26.4 or later.
2. The latest ChatGPT app.
3. A vehicle or head unit that supports CarPlay.

Steps:

1. Open CarPlay.
2. Select ChatGPT.
3. Select New voice chat.
4. Start speaking.

Important limitations:

• CarPlay conversations cannot access maps.
• They cannot access vehicle information.
• They cannot access live location.

Treat it as a conversational interface, not a navigation assistant. Use it to explain concepts, draft messages, summarize ideas, or think through problems. Use your car’s built-in systems for navigation and vehicle data.

Practical ways to get better results

Use GPT-Live differently from traditional voice assistants:

• Interrupt early. If the answer is going in the wrong direction, correct it mid-sentence instead of waiting.
• Use natural speech. You do not need rigid prompt formatting for basic tasks.
• Escalate hard topics. For complex reasoning, switch to Medium or High if available.
• Use it for live translation. Full-duplex interaction makes conversations feel less like walkie-talkie turns.
• Keep quick tasks on the default model. For facts, reminders, short explanations, and hands-free chores, the Instant-backed variants are faster.

For more on the architecture, model variants, and delegation design, see what is GPT-Live.

If you are a developer waiting to build directly with GPT-Live models, API access is not available yet. See what to use today instead. Teams building voice agents on OpenAI’s Realtime API can use Apidog for WebSocket session testing and backend mocking while GPT-Live API access remains unavailable. You can also download it free to keep voice integrations testable.

FAQ

Is GPT-Live free to use?

Yes. GPT-Live is available on ChatGPT’s Free tier through GPT-Live-1 mini. Paid plans get GPT-Live-1.

How do I switch between GPT-Live and Advanced Voice Mode?

Use ChatGPT’s voice settings. Advanced Voice Mode remains available and is currently the option to use if you need video or screen sharing.

Does GPT-Live work in CarPlay?

Yes. ChatGPT in CarPlay works on all plans with iOS 26.4 or later. Open ChatGPT from the CarPlay screen and start a new voice chat. Maps, vehicle data, and live location are not available there.

What are GPT-Live-1 Medium and High?

They are GPT-Live variants backed by GPT-5.5 Thinking at medium and high reasoning effort. They are slower than the default model but better for harder conversations.

Why does my ChatGPT not have GPT-Live yet?

The rollout is staged. Update the app and wait for the feature to reach your account. There is nothing to enable manually.