ChatGPT retires Advanced and Standard voice on Dec 11

What changes on December 11

OpenAI is retiring Advanced and Standard voice from ChatGPT on December 11, 2026. Anyone who still has either mode selected under Settings > Voice switches to Live automatically. Until that date, both older modes stay selectable. OpenAI started notifying people on November 10.

The company posted it on the OpenAI Newsroom account ↗ at 22:07 UTC on November 10, and published the same cutoff in a Help Center article ↗. The November 10 release notes ↗ repeat it. Devdiscourse ↗, citing ANI, reported the cutoff on November 11.

OpenAI's reason is usage. GPT-Live arrived in July. Since then, the company says, "voice usage has nearly doubled, and virtually all people using voice in ChatGPT now use GPT-Live." OpenAI did not publish the counts.

What Live, Advanced, and Standard actually are

Live is the mode that stays. The ChatGPT Voice article ↗, updated November 11, says Live runs on GPT-Live-1.1 or GPT-Live-1.1 Quip, depending on the plan. It is full-duplex: it listens and speaks at the same time, and it can hand a hard question to a frontier text model while the conversation continues. It uses web search, memory, plugins, connected apps, and text or images in the same chat.

Advanced was the previous real-time voice. It processed audio natively and waited until you stopped speaking. Standard chained speech-to-text, a language model, and text-to-speech. OpenAI says that path was slower and lost information between voice and text. It is also the path some people want to keep, because the answer follows the text model.

Live does not do video or screen sharing. Those stay on Advanced, in the iOS and Android apps, "until GPT-Live has this capability." OpenAI has not dated that handoff. December 11 removes Advanced as a voice you can pick for a normal call. It leaves Advanced running underneath video and screen share.

Those rows are from the Voice article, measured over a rolling 24 hours. Business Standard is 3 hours of GPT-Live-1.1, then 1.25 credits per extra minute. Enterprise on usage-based USD pricing is listed at $0.05 per minute. ChatGPT notifies you at the limit.

Old Advanced and Standard chats stay in history. You can continue them in text or in voice. After December 11, voice on those chats uses GPT-Live.

Custom GPTs retire the same day for most workspaces

The voice FAQ points at a separate retirement, and OpenAI's own page confirms the date. The custom GPT retirement FAQ ↗ schedules custom GPTs for December 11, 2026. FedRAMP workspaces, and Enterprise workspaces with a qualifying approved deferral, retire on February 11, 2027. That later date is not the default.

Until then, voice inside a custom GPT keeps using Advanced. On the retirement date the GPT and its page become inaccessible. When "Migrate to plugin" shows up under My GPTs, instructions become a skill, knowledge files copy over, and connected apps move. Custom actions do not. The original GPT turns read-only after migration, and the new plugin starts private. Old chats stay. Personal accounts (Free, Go, Plus, and Pro) already cannot create new GPTs, and the migrate button is not on every workspace today.

What people are saying

The argument is about losing a choice, and about a camera feature that is "retired" and still on duty.

"I'll be honest: I wish Standard were staying as an option. For some of us, its text-like depth and rhythm make moving between text and voice nearly seamless," wrote Storm and Flame ↗. They want to leave a long text thread, tap Voice, and keep the same depth.

Huzaifa ↗ put the camera exception in one line: "so Advanced voice is 'retiring' Dec 11 but it's still the thing powering live video and screen share. retired, still on shift." Kan Decods ↗ wrote that "the real test is whether GPT-Live can offer the same flexibility and even more."

Reddit was not reachable while this was reported, so these reactions are from X.

Our take

If Standard is why you use voice, use the next month. Both modes are still under Settings > Voice. Run the same prompt in each and judge the depth yourself. OpenAI says it is carrying Standard and Advanced habits into Live, naming "the depth and detail of responses, maintaining continuity across conversations, and better personalization." That work is named. A before-and-after is not published.

If you share your camera or your screen, leave that workflow in place on December 11. Advanced keeps doing that job until Live can. The voice article ↗ says Live does not support video or screen sharing, and that both remain on Advanced in the mobile apps.

If you maintain a custom GPT, treat December 11 as the deadline unless the workspace notice says February 11, 2027. Finish edits, migrate when the button is there, and rebuild any custom action as an app or an MCP server. The replacement plugin does not inherit your audience.

People already on Live for ordinary conversation can ignore the mode name. OpenAI says that is nearly everyone using voice. Plus is still 3 hours a day of GPT-Live-1.1, and that cap matters more than the picker.

If Live's style is the problem, look at Gemini Live, Meta Muse voice, and Cursor's voice chat for coding agents. ChatGPT Dictation, for a recording you edit before sending, is outside this retirement. Broader lists: AI voice generators and ChatGPT alternatives.