Your ChatGPT Advanced Voice Goes Away on December 11

If you talk to ChatGPT while you cook, walk or work through an idea, the voice mode you reach for may be about to change. OpenAI will retire ChatGPT Advanced Voice and Standard Voice on December 11, 2026, according to Interesting Engineering and Open Magazine. If you still use either mode, ChatGPT will switch you to GPT-Live automatically.

Your old voice chats will still be in your history, ready for you to pick up by typing or talking. And if you use your phone’s camera or share your screen during a call, Advanced Voice will keep handling those features for now. GPT-Live can work with still images you attach, and OpenAI plans to add live video and screen sharing later.

So what will feel different the next time you tap the waveform? Start with what changes on December 11 and what your plan gives you.

What Changes for ChatGPT Advanced Voice on December 11

Open the Voice menu today and, depending on your account, you may see three choices called Live, Advanced and Standard. Advanced and Standard will leave the picker on December 11, and ChatGPT will select Live for you if you still use either one. You can keep your current choice until the cutoff, and your chats carry over automatically. The timeline shows how ChatGPT got from its first spoken replies to one default voice mode.

A Quick Timeline of ChatGPT Voice

Why retire both now? OpenAI says almost everyone who uses ChatGPT Voice has already moved to GPT-Live, and voice usage has nearly doubled since GPT-Live arrived in July.

How GPT-Live Talks Differently From the Old Modes

Think about the last time you paused halfway through a sentence to find the right word. Advanced Voice could take that silence as its cue to answer, while GPT-Live can keep listening and give you time to finish. Here’s how that difference plays out across the three modes.

Listening and Speaking at the Same Time

GPT-Live uses full-duplex audio, which means you can interrupt it while it’s still talking. You might hear a quick “mhmm” while you explain something, or get a moment of silence while you think. You can ask GPT-Live to speak more slowly or give you a moment to finish.

GPT-Live is designed to pick out your voice amid background noise. A second person talking nearby can still cause an interruption, so headphones or a quieter room may help if that keeps happening.

Handing Hard Questions to a Bigger Model

Ask a question that needs a web search or more reasoning, and GPT-Live can send the work to another model while keeping the conversation going. At launch, GPT-5.5 handled that work in the background. If you ask about the weather or a game score, you may see a card with the details as you listen, and you can bring an image into the conversation or use ChatGPT’s memory.

Instant is the default reasoning level. Medium and High use GPT-5.5 Thinking and may add a second or two to a reply. You’ll need a paid plan with GPT-Live-1 to use those two levels.

Which Plans Get GPT-Live and How Much You Can Use

You can talk to GPT-Live on a Free account, but how long you can keep talking depends on your plan. ChatGPT counts daily Voice in Chat allowances over a rolling 24-hour period. Check your plan below for the model, included time and any per-minute charges. Deciding between plans? Our ChatGPT Go vs Plus comparison breaks down the price gap.

OpenAI’s July launch announcement listed GPT-Live-1 as the default for Go. The current Help Center lists GPT-Live-1 mini and gives Go users three hours. The Help Center also lists Business overages at one credit per minute, while MIXED News reported a rate of 1.25 credits per minute for Business.

You can use Voice in Chat on iOS, Android and the web. ChatGPT is also available through Apple CarPlay on supported iPhones. If your workspace is eligible, you can use Voice in Desktop with paired remote access in the macOS or Windows app and check OpenAI’s separate page for desktop usage limits.

What Happens to Video, Screen Sharing and Your Older Chats

If you use ChatGPT to look at something through your camera or talk you through an app, December 11 works a little differently for you. Advanced Voice will keep handling those sessions, and you’ll still be able to open chats you started in either older mode. Here’s where to find those features and what happens to the audio you’ve shared.

Camera and Screen Sharing Stay on Advanced for Now

• Live video. If you’re an eligible subscriber using ChatGPT on iOS or Android, tap the camera button during a supported voice chat to share live video.
• Screen sharing. On those same mobile apps, open the more-options menu during a call and choose Share Screen.
• GPT-Live support. Advanced Voice will keep powering both features until GPT-Live can handle them. OpenAI has given no date for that change.

Your Voice Chat History and Custom GPTs

• Earlier conversations. Open a chat from your history and pick up where you left off by typing or speaking. If you speak, GPT-Live will handle the new voice replies.
• Voice in custom GPTs. Custom GPTs are scheduled to keep using Advanced Voice through the transition to December 11. OpenAI is separately helping creators move from custom GPTs to plugins. Some Enterprise workspaces and FedRAMP workspaces have a later cutoff for custom GPTs.

Voices, Personalities and Audio Data

• Preset voices. You can choose from nine voices, including Arbor, Cove, Ember and Juniper. Viola and Rio are additional options in Brazil. GPT-Live offers preset voices and has protections against imitating real people.
• Tone and pace. ChatGPT’s preset personalities don’t carry over to Live. You can ask GPT-Live to speak more slowly or use a different tone during the conversation.
• Recordings. ChatGPT keeps audio from Live and Advanced with the transcript for 30 days. The same retention period applies to video from Advanced. Standard deletes the audio after transcription unless you choose to share it.

How to Try GPT-Live Before the Cutoff

You can try GPT-Live now and keep your current voice mode available until December 11. Start with a conversation you’d normally have in ChatGPT Advanced Voice so you can hear how Live handles your pauses, follow-up questions and requests for more detail. Here’s how to start a call on your phone or computer. If voice won’t start at all, our ChatGPT voice mode troubleshooting guide can help.

1. Open ChatGPT. Update the app on your iPhone or Android phone, or open ChatGPT in your browser.
2. Choose Live. Open your account menu, go to Voice and select Live from the available modes.
3. Start a call. On your phone, tap the waveform icon in the message bar. On the web, look for the Voice icon in the prompt box. Allow microphone access if asked, pick a voice if this is your first call, then start talking.
4. Try a reasoning level. If your paid plan gives you GPT-Live-1, tap the gear icon during a call. Start with Instant, or try Medium or High for a question that needs more thought.
5. Set up Voice for your routine. Turn on Background conversations to keep a call going when you switch apps or lock your phone. On supported mobile versions, Start with Voice opens new or empty chats in voice mode. If you use a supported iPhone with CarPlay, look for Start automatically in CarPlay after you’ve used ChatGPT there once.

If You Depend on Advanced or Standard Voice

Maybe you use Advanced Voice to talk through your day, or Standard Voice because its spoken replies match what you’d get by typing. The time before December 11 gives you a chance to see how GPT-Live handles the conversations you rely on. Try these adjustments with the tasks you do most often. Google’s rival has its own quirks, covered in our Gemini Live fixes.

• Ask for more detail. If a reply feels too brief, ask GPT-Live to expand on a specific point. You can ask it to speak more slowly, too, though there’s no precise playback speed control.
• Keep the written reply handy. If you need the exact wording of a text answer, type your question and follow the reply in the chat. If you’d prefer to speak your prompt first, Dictation turns a recording into text you can edit before sending.
• Cut down on interruptions. Try headphones or move to a quieter room if GPT-Live keeps reacting to nearby sounds. On an iPhone, Voice Isolation may help.
• Keep using Advanced for visual help. If your iOS or Android subscription gives you camera and screen sharing, continue using Advanced Voice for those features while OpenAI works on adding them to GPT-Live.
• Check the plugin transition. If you build custom GPTs that people use with voice, review OpenAI’s move to plugins before December 11. Plugin support in Live was reported in September, though the Help Center still lists plugins as unsupported.
• Share accessibility feedback. If you rely on Voice for accessibility, tell OpenAI which part of your routine is harder in Live. OpenAI says feedback will guide work on response detail, continuity and personalization.

OpenAI reversed an earlier decision to retire Standard Voice in September 2025 after users objected. The current cutoff for both older modes is December 11, 2026.

The Bottom Line on ChatGPT’s Voice Switch

If you still use Advanced or Standard Voice, ChatGPT will move you to GPT-Live automatically. Your conversations will be there when you open your history, and eligible mobile subscribers can still turn to Advanced Voice for camera and screen sharing. Before December 11, check your plan’s voice allowance and give Live a try.

Pick a conversation you already know well. Ask a follow-up, pause when you need to think and see how Live responds when you ask for more detail. By the time the switch happens, you’ll know what to say to get a reply that works for you.

FAQs