ChatGPT Retires Advanced and Standard Voice Mode for GPT-Live Dec. 11

OpenAI is retiring ChatGPT's Advanced Voice Mode and Standard Voice on December 11, 2026, folding every remaining user over to GPT-Live, the full-duplex voice model it introduced in July. OpenAI notified users of the change on November 10, 2026, and says the two older modes will disappear from the voice picker in Settings > Voice on the retirement date, with anyone still using them automatically defaulted to GPT-Live. Existing voice conversations stay in ChatGPT's history and can be picked back up by text or voice.

What's actually changing

Until now, ChatGPT has offered three ways to talk to it out loud: the original Standard Voice (speech converted to text, answered, then read aloud by a separate text-to-speech engine), Advanced Voice Mode (a lower-latency, more expressive speech-to-speech model introduced in 2024), and GPT-Live, the full-duplex model OpenAI shipped in July 2026. After December 11, only GPT-Live remains as a selectable voice experience in the ChatGPT app. Standard Voice and Advanced Voice Mode stop appearing in the model picker, and OpenAI is moving everyone who had either one selected onto GPT-Live by default.

This is a retirement of the legacy interaction modes, not of your data. OpenAI's help center article on the change is explicit that past voice conversations remain in your ChatGPT history and can be continued in text or voice — nothing you've already said to ChatGPT disappears on December 11.

Why OpenAI is retiring the older voice modes now

OpenAI's stated reasoning is usage, not just age. According to OpenAI, voice usage has nearly doubled since GPT-Live's July introduction, and the company says virtually everyone still talking to ChatGPT by voice has already gravitated to GPT-Live rather than sticking with the older modes. Maintaining three parallel voice systems — one that's barely used, one that's a stopgap, and one that's where almost all the usage and engineering investment already sit — isn't a good use of effort. OpenAI frames the move as a way to simplify the voice experience and put its focus on improving GPT-Live, rather than continuing to support Standard Voice and Advanced Voice Mode as parallel options indefinitely.

It's also not OpenAI's first attempt at this kind of consolidation. The company tried to retire Standard Voice once before, in 2025, and reversed course after user pushback. This time the retirement is paired with a concrete deprecation date and a direct replacement that's already carrying the bulk of voice traffic, which is a different situation than asking people to give up a mode most of them were still actively using.

What GPT-Live actually does differently

GPT-Live is built around full-duplex conversation rather than the strict turn-taking that defined every previous ChatGPT voice mode. Standard Voice and Advanced Voice Mode both fundamentally wait: you speak, the model listens until you stop, and only then does it process a response and speak back. GPT-Live instead listens and responds continuously, the way a person does in a real conversation. The model is deciding, many times a second, whether to keep listening, start speaking, pause, let you interrupt, or hand off to a tool — rather than treating the conversation as a rigid sequence of "your turn, my turn."

In practice, that means GPT-Live can follow pauses and changes in pace the way a human listener would, acknowledge you briefly while you're still mid-sentence without fully taking over the conversation, and sit quietly while you think without that silence being mistaken for the end of your turn. For harder requests — the kind that benefit from careful reasoning, a web search, or multi-step work — GPT-Live can delegate that work to one of OpenAI's other models in the background and bring the answer back into the conversation, rather than forcing the entire exchange through a single lightweight voice model.

Two versions of GPT-Live exist. GPT-Live-1 is the default for paid ChatGPT plans (Go, Plus, and Pro), while GPT-Live-1 mini is the default for Free accounts. Both are full-duplex; the difference is in the underlying model capacity rather than the interaction style. GPT-Live's own system card describes both as trained to respond safely using the same infrastructure OpenAI relies on for its flagship models, including when a conversation is handed off to one of those flagship models mid-call.

Standard Voice vs. Advanced Voice Mode vs. GPT-Live

What's not yet supported on GPT-Live

GPT-Live's retirement of the older modes comes with two carve-outs, and both matter if you rely on them. Live video and screen-sharing inside ChatGPT voice conversations continue to run on Advanced Voice Mode, because GPT-Live doesn't yet support those capabilities — OpenAI hasn't given a date for when that changes, so Advanced Voice Mode sticks around specifically for those sessions even after it disappears from the general voice picker. Separately, voice conversations inside custom GPTs also keep using Advanced Voice Mode, but only until the same December 11 deprecation date; after that, custom GPT voice chats lose their current voice backbone along with everything else. OpenAI's Advanced Voice Mode FAQ covers the mechanics of the mode that's being wound down, for anyone who wants the fuller picture of what it could do before this transition.

If your workflow depends on talking to ChatGPT while sharing your screen, or on voice inside a specific custom GPT, it's worth testing GPT-Live against your use case before December 11 rather than assuming full feature parity on day one. There's no indication that OpenAI plans to bring Standard Voice's specific text-to-speech pipeline back once it's gone; the gap being filled is GPT-Live catching up on video, screen-sharing, and custom GPT support, not a return of the older architecture.

What happens to your existing chat history

Nothing you've already said in Standard Voice or Advanced Voice Mode is deleted. OpenAI's help center confirms that conversations remain in your ChatGPT history and can be continued in text or voice — the retirement affects which voice model you can pick going forward, not the record of past conversations. If you had a long-running voice thread in Advanced Voice Mode, you'll still be able to scroll back to it, read the transcript, and keep the conversation going by typing or by switching it over to GPT-Live.

How to check which voice mode you're on right now

You can see which mode you're using today in the ChatGPT app under Settings > Voice. If Advanced Voice Mode or Standard Voice is still selected there, it will keep working as normal until December 11; after that date, the setting switches automatically to GPT-Live and the older options are no longer listed as choices. There's nothing you need to do in advance to keep your account working — the switch happens on OpenAI's side — but if you want to get used to GPT-Live's different conversational rhythm before it becomes your only option, switching over now is the straightforward way to do that. OpenAI's general voice mode FAQ is the reference point for how the picker behaves and which plans get which default, and it's the page likely to be updated first if OpenAI adjusts any part of this rollout.

Does this change what free users get versus paid plans?

Access to GPT-Live itself isn't new or tied to this retirement — it's been rolling out as the default voice experience since July. What changes on December 11 is simply that it stops being one option among several and becomes the only one. Free ChatGPT accounts default to GPT-Live-1 mini, while Go, Plus, and Pro subscribers default to the larger GPT-Live-1. Neither tier gets a separate charge for voice as part of this transition; it's a replacement of the underlying model behind an existing feature, not a new paid add-on. Anyone searching for standalone "GPT-Live pricing" is more likely thinking of the developer-facing API version of the model, which OpenAI has discussed separately for building voice features outside of ChatGPT — that's a distinct product from the in-app ChatGPT voice experience this retirement affects.

Why some users are unhappy about losing Standard Voice

Not everyone is pleased with the change. In the run-up to the December 11 date, some ChatGPT users have been vocal on OpenAI's own community forum about preferring Standard Voice's response style, arguing it tends to give more detailed, deliberate answers than the faster, more conversational tone GPT-Live defaults to. That tension — between a model built to feel like natural back-and-forth conversation and users who wanted voice mode to behave more like a spoken-aloud version of a careful written answer — is the same friction that derailed OpenAI's first attempt to retire Standard Voice back in 2025. Whether the same pushback reverses this retirement, too, is an open question; this time, the replacement is already handling the overwhelming majority of voice traffic, which gives OpenAI less reason to walk it back.

Part of a broader pattern of model consolidation

The voice mode retirement follows a shape OpenAI has repeated elsewhere this year: ship a specialized model, let it absorb usage from older defaults, then formally sunset what it replaced. It's the same arc behind OpenAI's rollout of GPT-6 Cyber as a dedicated security model rather than folding those capabilities into a general-purpose release, and it echoes how OpenAI has positioned its broader GPT-5.2 model family as the reasoning engine that specialized products like GPT-Live quietly hand work off to. Other AI labs are iterating on a similar cadence — Google's jump from Nano Banana 2 to Nano Banana Pro is a comparable example of a fast-moving product line replacing its own predecessor within months rather than years.

What's next

Between now and December 11, nothing changes for anyone still using Standard Voice or Advanced Voice Mode — both keep working exactly as they do today. The practical deadline is the date itself: after December 11, 2026, GPT-Live is the only voice mode available in ChatGPT for everyone except the two carved-out cases (video/screen-sharing and custom GPT voice chats, the latter only until that same date). OpenAI hasn't published a timeline for when GPT-Live will pick up video and screen-sharing support, which is the detail most likely to determine how smooth this transition feels for anyone who relied on Advanced Voice Mode for more than plain conversation. Expect OpenAI to keep updating its voice mode help documentation as GPT-Live closes those remaining gaps, and expect continued debate from users over whether GPT-Live's conversational style is a genuine upgrade or just a different trade-off from the modes it's replacing.

Written byTheo Park

Theo Park runs the AI desk at Pandromeda. He follows model launches from the frontier labs and the open-weight community, tracks the assistants and developer tools built on them, and explains what each release changes on pricing, capability and safety. His reporting leans on primary sources: model cards, technical reports, API documentation and the companies' own announcements.

More from Theo ParkAll AI coverage