AI

OpenAI will retire ChatGPT's Advanced and Standard voice on Dec. 11, 2026

GPT-Live becomes the default voice, but video, screen sharing and custom GPTs complicate the picture, and some users are pushing back.

11 min read
Hand holding a smartphone with glowing red, green and purple waveform lines flowing across a black background.
Image by geralt from Pixabay

OpenAI's release notes say Advanced and Standard voice leave ChatGPT's voice picker on December 11, 2026, with GPT-Live as the replacement. Live video, screen sharing and custom GPTs are exceptions with unclear timelines, and the company's claim that voice usage nearly doubled rests on its own word.

OpenAI plans to remove Advanced voice and Standard voice from ChatGPT on December 11, 2026, according to its own release notes. After that date, GPT-Live, the voice model the company introduced in July, becomes the voice experience for nearly everyone. Anyone who had picked one of the older modes is moved to Live automatically.

The change matters because voice is one of the main ways people talk to ChatGPT, and the older modes have loyal users who say Live does not feel the same. It also matters because the consolidation is less complete than the headline suggests: Advanced voice keeps running live video and screen sharing, and the schedule for custom GPTs is unclear. This is also not the first attempt. OpenAI announced the retirement of Standard Voice in September 2025 and then backed off.

What OpenAI has announced

The date and where it comes from

Early coverage of the retirement leaned heavily on blog posts, which raised the question of whether the December 11 date was solid. It is. OpenAI's ChatGPT release notes carry an entry dated November 10, 2026, titled "Retiring Advanced and Standard Voice," that sets the cutoff. The same text appears on the release notes page at openai.com, and OpenAI has a dedicated Help Center article on the change.

OpenAI's newsroom account on X posted the announcement on November 10 as well. Those three channels are all controlled by OpenAI, so they are not independent of one another, but they come from the party that is actually making the decision, which is the strongest kind of source for a product change. Outlets such as Android Authority, Neowin, Dataconomy and TechRepublic restate OpenAI's notice; they add context and user reaction but do not independently confirm anything beyond what the company said.

We’re retiring Advanced and Standard voice from ChatGPT Voice on December 11, 2026.

OpenAI Newsroom on X, November 10, 2026

The notice period works out to 31 days, from November 10 to December 11. December 11, 2026 falls on a Friday. OpenAI's text does not give a time of day or a time zone for the cutoff.

What happens to existing chats

According to OpenAI's Help Center article, the transition has three stages. Users were notified on November 10. Until December 11, they can keep using either older mode. After that, Advanced and Standard disappear from the model picker under Settings > Voice, and anyone who had selected them is defaulted to Live voice.

Past conversations are not deleted by the change. The same help article says they remain in ChatGPT history and can be continued in text or voice, with voice defaulting to GPT-Live once the older modes are gone.

What is not going away yet

Live video and screen sharing

The phrase "going away" overstates the case in one important respect. OpenAI's release notes say live video and screen sharing will continue to be powered by Advanced voice until GPT-Live gains that capability. In practice, a version of Advanced voice stays running behind those features even after it vanishes from the picker.

No date has been given for when GPT-Live will take over those features. Android Authority and Analytics Insight both noted that OpenAI offered no timeframe. Anyone who relies on talking to ChatGPT while showing it a camera feed or a screen therefore has no announced deadline to plan around.

Custom GPTs: an open question

OpenAI's help article says voice with custom GPTs will continue using Advanced voice until its deprecation on December 11, 2026. That sentence does not say what custom-GPT voice users get afterward. Interesting Engineering reads it as meaning custom GPTs will also lose access to Advanced voice on that date, but that is the outlet's interpretation, not an OpenAI statement.

The reasonable conclusion is that custom-GPT voice is on an unclear schedule. OpenAI may clarify it before or on the cutoff date, so the help article is worth rereading closer to December 11.

A separate, earlier change on the Mac

Voice has been shuffled before on desktop. OpenAI's release notes say it retired the Voice experience in the ChatGPT macOS app on January 15, 2026, while voice stayed on chatgpt.com, iOS, Android and the Windows app. A later Help Center page, however, describes voice in the macOS desktop app for Work and Codex, so the current state on the Mac is not clear from the sources reviewed.

What GPT-Live is and how it differs

TechCrunch reported on July 8, 2026 that OpenAI had released two models, GPT-Live-1 and GPT-Live-1 mini. Both are described as full-duplex, meaning they can listen and speak at the same time. That allows a user to interrupt naturally rather than waiting for the assistant to finish, and TechCrunch said the models enable features such as live translation.

Secondary outlets add details that trace back to OpenAI's launch and that the research for this article could not check against the original post. VoiceOS, a vendor blog, says GPT-Live adds conversational cues such as "mhmm" and "yeah," handles interruptions, and passes hard questions to a stronger model in the background, which it names as GPT-5.5 at launch. Treat those specifics as reported, not confirmed.

Three modes, three designs

OpenAI's Help Center explains the technical differences. Standard voice chained three models: speech-to-text, a large language model and text-to-speech. OpenAI says that design made responses slow and lost information between voice and text. Advanced voice ran on what OpenAI calls a previous turn-based voice model, and Live voice is powered by GPT-Live, which the company describes as combining natural conversation with frontier intelligence.

ChatGPT's three voice modes and their status as of November 2026
Mode Architecture (per OpenAI and TechCrunch) Turn-taking Status
Standard Three chained models: speech-to-text, a large language model, text-to-speech Turn-based Removed from picker Dec. 11, 2026
Advanced Previous turn-based voice model (processes audio natively, per Dataconomy) Turn-based Removed from picker Dec. 11, 2026; still runs live video and screen sharing
Live (GPT-Live) Full-duplex (per TechCrunch) Listens and speaks at once Default and only selectable voice after Dec. 11, 2026

Context: where Advanced voice came from

Wikipedia's entry on GPT-4o says Advanced Voice Mode reached ChatGPT Plus and Team subscribers in September 2024 and used GPT-4o's native audio handling. OpenAI later retired GPT-4o from ChatGPT on February 13, 2026, but its help page said ChatGPT Voice was not changing as part of that update. The voice product has therefore outlived the model it launched on.

What Live does differently for users

OpenAI's voice FAQ says the preset ChatGPT personalities do not currently apply to Live. Users can still ask for a different tone or pace within a conversation. Another help page describes Live voice in Work on web, iOS and Android, so it is spreading beyond the basic chat screen.

Which plans received GPT-Live at launch is disputed. A blog on Build Fast With AI says GPT-Live-1 mini brings the full-duplex design to Free-plan users and that all plans get GPT-Live, while MLQ.ai's account of the launch differs. The research did not retrieve OpenAI's own launch post, so plan availability at launch remains unverified here. Whatever the launch picture, OpenAI's retirement notice says virtually all people using voice are already on GPT-Live.

OpenAI's case: usage nearly doubled

OpenAI's stated reasoning is that adoption has already made the older modes redundant. Its release notes say voice usage has "nearly doubled" since GPT-Live launched in July and that virtually all people using voice in ChatGPT now use GPT-Live. The company says retiring the older modes lets it simplify the voice experience and concentrate on improving GPT-Live.

The usage figure is a single-source claim from the company, and it comes with no absolute numbers, no baseline period and no definition of what counts as usage. A doubling could mean twice as many people, twice as many sessions, or twice as much talk time. For that reason it cannot be charted or independently weighed.

Reasons for caution about the number

  • Several secondary outlets say Live replaced Advanced as the default on July 8, which could shift much of the usage mechanically. That point was not verified against OpenAI's launch post.
  • New capabilities such as full-duplex conversation and live translation, along with a wider rollout, could raise usage regardless of whether people prefer the new voice.
  • The statistic does not reveal whether the heaviest voice users, who are often the loudest critics, are among the "virtually all" now on Live.

The release notes also contain a quiet concession. OpenAI says it will carry lessons from the old modes into GPT-Live, naming the depth and detail of responses, continuity across conversations, and personalization. Read plainly, that list points to areas where Live is not yet where the older modes were, though OpenAI does not say so in those terms. Neowin goes further and reports that OpenAI acknowledged some users may prefer the response style of Advanced or Standard voice; that wording was not found in the OpenAI text reviewed, so it is Neowin's characterization.

Backlash, and the September 2025 precedent

TechRadar reports that a vocal group of users is unhappy, with complaints appearing on Reddit and X. According to a syndicated version of that coverage, the complaints are that Live fills gaps with filler such as "yeah" and "anyway," interrupts and breaks a train of thought, and shifts its tone more than the older modes did. YourStory and other outlets report that selecting the same named voice in Live does not recreate the experience people had with Standard or Advanced.

A post on OpenAI's own community forum asks the company to keep Standard as an option. These accounts are self-selected anecdotes, not a survey. Complaints clearly exist, but no data show how many users share them, and the usage claim above suggests that many people have adopted Live without objection, though that too is OpenAI's account.

A repeat of 2025?

TechRadar reports that in September 2025 OpenAI announced Standard Voice would be retired within 30 days, then reversed. A Yahoo Tech article quotes OpenAI's message at the time as saying Standard Voice would stay available while the company addressed feedback on Advanced Voice. The reversal came on September 9, 2025, according to the timeline in the research.

The precedent is real, but the circumstances differ. The 2026 announcement rests on a different rationale: by OpenAI's account, nearly everyone using voice has already moved to Live, and Advanced voice is being retired alongside Standard rather than offered as the fallback. Commentators at Neowin and Digital Today have speculated about another reversal, but that is opinion, not evidence. Whether OpenAI holds to December 11 is unknown.

What it means and what to watch

The retirement is a bet that one voice model, built to listen and talk at once, can replace a set of modes that each had a different feel. For many people that may be an upgrade. For others, the pacing, tone and depth they chose deliberately will change, and OpenAI's own notes concede that some qualities of the older modes still need to be rebuilt in GPT-Live.

Three things are worth watching before and after December 11: whether OpenAI clarifies the custom-GPT wording, whether it offers any timeline for moving video and screen sharing to GPT-Live, and whether the date itself holds, since the release notes page is updated over time. The strongest evidence on whether Live is good enough will not come from OpenAI's usage claim but from whether complaints fade once the older modes are no longer an option.

Frequently asked questions

When do ChatGPT's Advanced and Standard voice end?

OpenAI's release notes set the retirement for December 11, 2026. From that day the two modes can no longer be selected under Settings > Voice, and users who had chosen them are moved to Live voice. OpenAI has not stated a time of day or time zone for the cutoff.

What is GPT-Live?

GPT-Live is OpenAI's voice model, introduced in July 2026. TechCrunch reported that GPT-Live-1 and GPT-Live-1 mini are full-duplex models that can speak and listen at the same time, so users can interrupt naturally. OpenAI describes the model as combining natural conversation with frontier intelligence.

Will old voice conversations be deleted?

According to OpenAI's Help Center, conversations stay in ChatGPT history and can be continued in text or voice. After the cutoff, voice defaults to GPT-Live. The sources reviewed conflict on how long audio clips are retained, so that question is unresolved.

Will live video and screen sharing still work after December 11?

OpenAI says those features will keep running on Advanced voice until GPT-Live can do the same. The company has given no timeline for that move, according to Android Authority and Analytics Insight.

Could OpenAI reverse the decision as it did in 2025?

It is possible but unknown. In September 2025 OpenAI announced Standard Voice would be retired within 30 days and then kept it available while it addressed feedback. The 2026 plan rests on OpenAI's statement that virtually all voice users are already on GPT-Live, and no source reviewed indicates a reversal.

Sources (24)

  1. ChatGPT release notes , OpenAI Help Center (2026-11-10)
  2. Release Notes , OpenAI
  3. Retiring Advanced and Standard Voice in ChatGPT , OpenAI Help Center
  4. OpenAI Newsroom on X: retiring Advanced and Standard voice , OpenAI Newsroom (X) (2026-11-10)
  5. ChatGPT Voice , OpenAI Help Center
  6. Retiring GPT-4o and other ChatGPT models , OpenAI Help Center
  7. ChatGPT Work and Codex , OpenAI Help Center
  8. Getting started with ChatGPT , OpenAI Academy
  9. OpenAI releases new voice models for more natural live conversations , TechCrunch (2026-07-08)
  10. 'I'm beyond disappointed because ChatGPT means a lot to me': Users are furious that OpenAI is retiring their favorite voice modes , TechRadar
  11. OpenAI is retiring ChatGPT's Advanced and Standard Voice modes, sparking user backlash , Neowin
  12. OpenAI is sunsetting ChatGPT Advanced and Standard voice next month , Android Authority
  13. ChatGPT Is Retiring Two Voice Modes Dec. 11 — What Happens to Your Chats? , TechRepublic
  14. OpenAI To Retire ChatGPT Advanced And Standard Voice On December 11 , Dataconomy (2026-11-12)
  15. ChatGPT Voice Overhaul: OpenAI to Replace Advanced Modes with Live , Analytics Insight
  16. OpenAI pulls the plug on two major voice modes as GPT-Live takes over , Interesting Engineering (2026-11-11)
  17. OpenAI retires older ChatGPT voices for chatty Live mode , YourStory
  18. Users are furious that OpenAI is retiring their favorite voice modes , Wilson's Media
  19. GPT-Live Is Here: ChatGPT's Voice Now Listens and Talks at Once, Just Like 'Her' and JARVIS , VoiceOS
  20. GPT-Live Review: OpenAI's Full-Duplex Voice Model Explained (July 2026) , Build Fast With AI
  21. OpenAI Launches GPT-Live-1, a Full-Duplex Voice Model That Listens and Speaks Simultaneously , MLQ.ai (2026-07-09)
  22. Feedback Regarding the Retirement of Standard and Advanced Voice , OpenAI Developer Community
  23. ChatGPT's Standard Voice Mode was nearly gone , Yahoo Tech (2025-09-09)
  24. GPT-4o , Wikipedia

Share this article

Back to list