Retirement of ChatGPT’s Advanced and Standard voice modes set for December 11

On December 11, 2026, ChatGPT’s Advanced and Standard voice modes will give way to Live. Video and screen sharing remain an exception.

Two options are set to disappear from ChatGPT’s voice model picker: Advanced and Standard. The retirement notice in the Help Center sets the date for December 11, 2026. Users who selected either mode will automatically switch to Live. Until then, both options remain available under Settings > Voice.

The company attributes the decision to the adoption of GPT-Live, introduced in July. According to its statements, voice usage has nearly doubled since launch, and virtually everyone using voice in ChatGPT now uses that model. No precise figures or measurement details accompany these claims. Focusing on Live is intended to simplify the offering and concentrate development on that experience. Three ways to process speech

The retiring modes use two different architectures. In Standard, speech is first transcribed into text, a language model prepares a response, and a third model converts it into speech. In Advanced, a single model processes and generates audio directly, but the exchange still follows distinct turns: it waits for the user to finish speaking before responding.

The architecture described in the GPT-Live introduction is full-duplex, allowing simultaneous listening and speech generation. The model can offer brief acknowledgments, wait through a hesitation, or respond to an interruption. It can also delegate a complex question to a text model and continue the conversation while that task runs in the background.

The transition also affects response style. The company acknowledges that some people prefer how Advanced or Standard responds. Greater response depth, continuity across conversations, and personalization are among the announced areas of improvement for Live; the notice does not promise to reproduce the previous modes’ behavior exactly. Advanced remains necessary for some uses

Live video and screen sharing will continue to run on Advanced until GPT-Live supports those capabilities. The current ChatGPT Voice guide confirms that Live does not yet offer them. The retirement notice provides no date for transferring these capabilities.

A related timeline applies to custom GPTs, whose voice conversations will continue to use Advanced until those GPTs are retired. The custom GPT migration FAQ sets the general retirement date for December 11, 2026, with different deadlines for certain qualifying enterprise workspaces. It provides for migration to plugins, depending on the options available to the account and workspace.

Existing conversations with Advanced and Standard will remain in ChatGPT history. They can be continued in text or voice; voice will use GPT-Live after the transition. Keeping the conversation thread preserves access to earlier exchanges, even as the model and its response style change.