OpenAI will discontinue two legacy voice modes for ChatGPT next year, with GPT-Live becoming the default voice experience.

·by Henderson·Engineering
OpenAI will discontinue two legacy voice modes for ChatGPT next year, with GPT-Live becoming the default voice experience.
Key Points
  • OpenAI will discontinue two legacy voice options on December 11.
  • GPT-Live will become the new default voice experience, significantly enhancing conversational fluency.
  • Users need to adjust before the deadline, or they will be automatically switched to GPT-Live.
  • The features of advanced voice and standard voice will still be available during the transition period.

OpenAI will stop offering two legacy voice options next month, at which point GPT-Live will become the platform's default voice experience. The advanced voice and standard voice will be removed from the model selector on December 11, 2026. This change will affect users who still prefer the old conversation modes. OpenAI stated that since its launch in July, nearly all ChatGPT voice users have transitioned to using GPT-Live, and during this period, voice usage has nearly doubled, indicating a significant shift towards the new system.

Users will still be able to use advanced voice and standard voice during the transition period, but those relying on these two options need to make adjustments before the December deadline. This retirement will also affect custom GPTs currently using the legacy voice technology.

New Features of GPT-Live

GPT-Live features a different architecture compared to its predecessors. The model can listen and speak simultaneously, allowing conversations to flow naturally without the constraints of traditional turn-taking. This capability enables quicker exchanges and allows for natural interruptions. It can also respond to users with brief replies while they are speaking. When someone needs time to think, the model can remain silent, preventing the conversation from feeling stiff. Additionally, GPT-Live can connect with OpenAI's text-based models to handle more complex requests.

It can delegate complex questions and multi-step tasks without ending the current voice conversation, allowing users to continue talking while the system processes more complicated work.

Impact of Legacy Voices

The advanced voice employs a more traditional approach. Its underlying model directly processes incoming and outgoing audio but must wait for the user to finish speaking before generating a response. The standard voice relies on three separate models, one that converts speech to text, another that generates a response, and a third that converts that response back into audio. These additional processing stages often lead to delays and information loss between audio and text. OpenAI now aims to integrate these functionalities into a single voice experience. The company also plans to improve response depth, conversational coherence, and personalization during the development of GPT-Live.

OpenAI began notifying users about this retirement on November 10. Existing advanced voice and standard voice users can maintain their preferred settings until December 11. After that date, these two options will disappear from the voice settings. ChatGPT will automatically switch affected users to GPT-Live, requiring no manual migration of existing conversations or initiation of new chats. Previous conversations will remain accessible through ChatGPT's history, and users can continue these discussions via text or the new Live voice experience.

Custom GPTs will also lose access to advanced voice on December 11. Conversations with custom GPTs will continue to use the legacy model until the retirement date. Developers and users should consider this change when planning existing voice workflows.

An important feature will still retain the current voice infrastructure. According to OpenAI, real-time video and screen sharing will continue to operate through advanced voice. GPT-Live currently does not support these features. The company will maintain the old system until its new model achieves equivalent capabilities. This leaves a temporary distinction between standard voice conversations in ChatGPT and its interactive visual features. During the transition period, users will still have access to real-time video and screen sharing. OpenAI has not provided a specific timeline for introducing these capabilities into GPT-Live.

Currently, December 11 marks the end of these two optional legacy voice modes, but not all functionalities driven by advanced voice.

Advantages of GPT-Live and the Retirement of Legacy Voices

This change by OpenAI signifies a significant upgrade in voice technology. The design of GPT-Live allows users to enjoy a more natural conversational experience, which is crucial for enhancing the fluency and efficiency of user interactions. With the retirement of legacy voice modes, users will need to adapt to the new system, which may impact those reliant on the old modes. Nevertheless, OpenAI has retained some features of advanced voice to ensure a smooth transition period. This strategy not only demonstrates OpenAI's commitment to user needs but also reflects its ongoing innovation in voice technology.

H
About the author
Henderson