What to know
- ChatGPT voice mode is now available within the main chat, not as a separate interface.
- The update supports real-time visuals, including images and maps, during voice chats.
- Voice Mode is still using GPT 4o, instead of GPT 5.1.
- Voice mode is rolling out to all users on both mobile and web platforms.
- Users must update their ChatGPT app to access the new voice capabilities.
The new ChatGPT update allows users to activate voice mode directly from the main chat window, eliminating the need to switch to a dedicated voice interface. Once voice mode is enabled, conversations can seamlessly blend speech and text, letting users speak naturally while viewing responses, reviewing past messages, and accessing visual aids like images or maps.

Key features of the new ChatGPT Voice mode
- Instant live transcripts of voice chats are now available.
- Visual outputs such as maps and images appear in real time during voice conversations, a feature previously unavailable in the old voice-only mode.
- The system still uses GPT-4o, instead of the latest GPT-5.1, for voice replies.
- Paid plans begin with GPT-4o for voice sessions, switching to GPT-4o mini when GPT-4o minutes are exhausted. Free users start with GPT-4o.
As mentioned on X by OpenAI, the new voice mode is available to all ChatGPT users on both web and mobile apps, regardless of region. Access requires updating the ChatGPT app to the latest version. No additional subscription is needed for the core voice features, though premium plans offer advanced model usage.
How to revert to old ChatGPT Voice Interface
If you don't like ChatGPT Voice integrated into the main chat, you can keep the two separated as well. Here's how to go back to the old ChatGPT Voice interface:
- Tap on your profile icon in the bottom left corner.
- Select Voice.
- Turn on Separate Mode.

While Paid plans deliver "nearly unlimited" use of ChatGPT Voice with GPT-4o, free users will have the GPT-4o mini model with limited usage each day. It's a small but important asterisk that may put users off, especially since texts and voice, while being on the same chat, use different models. One can only hope Voice to catch up in future updates.
Discussion