Productive— faster every day
For your professionTeachersStudentsManagersMarketingDevelopersFreelancersParentsJournalists

AI & productivity · November 16, 2026 · 3 min read

ChatGPT retires Advanced and Standard voice on 11 December: only GPT-Live remains

OpenAI will retire the older Advanced and Standard voice modes on 11 December 2026. If you dictate tasks, notes or emails to ChatGPT, you have less than a month to try your routines in GPT-Live.

What happened: On 10 November 2026 OpenAI announced that it will retire the Advanced and Standard voice modes in ChatGPT on 11 December 2026. GPT-Live, which the company launched this summer, becomes the default and the only selectable voice model. Until then all three keep working. If you have one of the older modes selected, ChatGPT will move you to GPT-Live automatically after the deadline. Your existing voice conversations stay in your history.

OpenAI explains the change with numbers: voice usage has nearly doubled since GPT-Live launched, and according to the company virtually everyone already picks the new mode. It wants to focus development on a single model.

What it means for you: If you don't talk to ChatGPT, nothing changes. If you use voice for work, for example to dictate tasks on the way from a meeting, rehearse a presentation or control connected tools, which we covered in September, expect three things:

  • A different conversational rhythm. Advanced waited until you finished and then replied. GPT-Live holds a more fluid conversation and can be interrupted mid-sentence. When you dictate longer thoughts with pauses, it may jump in sooner than you expect.
  • A different delivery. According to user feedback, the same voice doesn't sound quite the same in GPT-Live as in the older modes. It is a matter of habit, but if you use voice to practise a language or a presentation, you will notice.
  • An exception for video and screen sharing. Live video and screen sharing will keep running on Advanced voice until GPT-Live supports them. OpenAI gave no timeline.

A test worth five minutes: dictate your usual instruction in GPT-Live, for example "summarize today's meetings from my calendar and, from the notes I'm about to dictate, prepare a draft email for the team". Compare the result with what you are used to.

How to get started:

  1. Open the voice settings in ChatGPT (Settings → Voice) and check which mode you have selected.
  2. Switch to GPT-Live now and use it for a week in your normal routine while you can still go back.
  3. Go through the routines you care about: dictating tasks, voice commands for connected tools, thinking out loud at length.
  4. If the interruptions bother you, try starting with "wait until I say done". It helps with longer dictation.
  5. Do you use voice in custom GPTs? Check OpenAI's help article for what changes for them after 11 December, and let the colleagues who use them know.

What to watch out for:

  • AI proposes, you approve. This applies twice over with voice. A misheard name or date is easy to miss in a dictated email. Keep actions that send or change something in confirmation mode, and read the result on screen.
  • Don't leave it to the last day. After 11 December the older modes will be gone from the menu and you won't be able to compare.
  • Feature availability may differ. OpenAI promises to bring the strengths of the older modes to GPT-Live over time. It hasn't said exactly what or when. Follow the current state in the ChatGPT release notes.

Source: OpenAI Help Center