Prepare for ChatGPT Voice Retirement on Dec 11

OpenAI will retire Advanced Voice and Standard Voice in ChatGPT on December 11, 2026. After that date, everyone who talks to ChatGPT uses GPT-Live. The company notified users on November 10. Voice usage nearly doubled after GPT-Live launched in July, and almost all voice sessions already run on the newer model.

This guide explains the exact timeline, the technical differences, what stays on Advanced Voice, and the steps to take before the automatic switch.

What changes on December 11

Advanced and Standard Voice disappear from the model picker under Settings > Voice. Anyone who had selected either mode moves to Live automatically. Existing conversations stay in your history. You can reopen them and continue by text or by voice. New voice replies use GPT-Live.

Custom GPT voice conversations also switch on the same day. Live video and screen sharing continue to run on Advanced Voice until GPT-Live supports those features. OpenAI has not set a date for that support.

You can still choose Advanced or Standard Voice until December 11. After that day the options are gone.

How GPT-Live differs from Advanced and Standard

GPT-Live is a full-duplex model. It listens and speaks at the same time. It can acknowledge you with short phrases such as “mhmm” or “yeah,” handle interruptions, and stay quiet while you think. When a question needs deeper reasoning or tools, it delegates the work to a frontier text model and keeps the conversation going.

Advanced Voice processed audio directly but waited for you to finish speaking before it replied. The result felt more like turn-based radio. Standard Voice chained three separate models: speech-to-text, a large language model, and text-to-speech. That path produced slower responses and lost nuance between voice and text.

OpenAI says it is carrying lessons from both older modes into GPT-Live. The company is improving response depth, continuity across conversations, and personalization.

Watch the official GPT-Live introduction

OpenAI released this short video when GPT-Live first powered ChatGPT Voice. It shows the full-duplex style and how the model handles natural back-and-forth.

Step-by-step: prepare before December 11

1. Update the ChatGPT app on iOS, Android, or the desktop client so you see the current voice options.
2. Open Settings > Voice. Note which mode is currently selected.
3. Switch to Live and run a few short conversations. Interrupt the model mid-sentence and observe how it responds.
4. If you rely on live video or screen sharing, keep Advanced selected for those sessions until GPT-Live supports them.
5. Review any custom GPTs you use with voice. Their voice path also moves on December 11.
6. Export important conversation history if you want a local copy. OpenAI provides a data export option in Settings.

Test Live in the environments where you normally talk to ChatGPT—while walking, in a quiet room, or with background noise. Note any differences in response style or speed so the automatic switch does not surprise you.

What stays the same after the switch

Past voice conversations remain in your ChatGPT history. You can open an old thread and continue it. If you speak, GPT-Live handles the new turns. Text replies continue to work exactly as before.

Live video and screen sharing stay on Advanced Voice for now. When GPT-Live gains those capabilities, OpenAI will update the experience. Until then, start a video or screen-share session and the older mode still powers the audio side.

Plan limits for GPT-Live hours do not change with this retirement. Check your current plan’s voice allowance in the app if you use Live for long sessions.

Tips for a smooth transition

Speak in shorter bursts at first if you are used to Advanced Voice’s turn-based style. GPT-Live can take interruptions, but clear pauses still help when you need a longer answer.

If a response feels thinner than the old Standard mode, ask a follow-up. GPT-Live can hand the harder part to a stronger text model while it keeps talking to you.

For multi-step or tool-heavy work, start the conversation in text, then switch to voice once the context is set. This matches the way GPT-Live delegates complex tasks.

Review privacy settings under Data controls before you talk about sensitive topics. The same account-level choices apply to voice as to text.

If you have not already explored how Voice interacts with stronger models, the earlier guide on using ChatGPT Voice with GPT-5.6 and Astra covers model selection inside spoken sessions.

Troubleshooting after December 11

I cannot find Advanced or Standard Voice. That is expected after December 11. Live is the only selectable option in Settings > Voice.

Video or screen sharing still works. Those features continue on Advanced Voice until GPT-Live supports them. No extra setting is required.

Old conversations look different. The history is intact. New voice replies simply use GPT-Live.

Live feels different from what I was used to. Give it a few sessions. OpenAI is incorporating feedback from the older modes into ongoing GPT-Live improvements.

Custom GPT voice stopped working the old way. Custom GPTs move to Live on the same date. Check the custom GPT retirement notes if the behavior surprises you.

Conclusion

December 11, 2026 is the hard cutoff for Advanced and Standard Voice in ChatGPT. The switch to GPT-Live is automatic for anyone still on the older modes. Live video and screen sharing remain the main exceptions that keep Advanced Voice active.

Try Live before the date if you still use Advanced or Standard. Confirm your Settings > Voice selection and review any custom GPTs you rely on. The change simplifies the product while OpenAI focuses development on the full-duplex model.

Sources

• Retiring Advanced and Standard Voice in ChatGPT — OpenAI Help Center
• Introducing GPT-Live — OpenAI announcement, July 8, 2026
• ChatGPT Voice — current voice options
• This is the new ChatGPT Voice, powered by GPT-Live — official OpenAI video