OpenAI will discontinue two legacy voice options in ChatGPT next month, leaving GPT-Live as the platform’s default voice experience. Advanced and Standard voice will disappear from the model picker on December 11, 2026.

The change affects users who still prefer the older conversational modes. OpenAI says nearly all ChatGPT voice users have already switched to GPT-Live since its introduction in July. Voice usage has also nearly doubled during that period, signaling a significant shift toward the newer system.

Users can continue accessing Advanced and Standard voice throughout the transition. However, those who rely on either option will need to adjust before the December deadline. The retirement also affects custom GPTs that currently use the older voice technology.

GPT-Live uses a different architecture from its predecessors. The model can listen and speak simultaneously, allowing conversations to flow without conventional turn-taking.

That capability enables quicker exchanges and more natural interruptions. It can also acknowledge users with brief responses while they speak. The model can remain silent when someone needs time to think, making conversations feel less rigid.

GPT-Live also connects with OpenAI’s text-based models for more demanding requests. It can delegate complicated questions and multi-step tasks without ending the ongoing voice conversation. This allows users to continue speaking while the system handles more complex work in the background.

Advanced voice followed a more traditional approach. Its underlying model processed incoming and outgoing audio directly. However, it had to wait until users finished speaking before generating a response.

Source link