OpenAI has improved ChatGPT’s voice mode

OpenAI has introduced an updated version of its ChatGPT voice mode, improving the expressiveness and naturalness of speech. The new mode, called Advanced Voice, became available over the weekend and now provides more subtle intonation, proper pauses, accents and emotional coloring of phrases, including sarcasm and empathy.

One of the key innovations is the continuous translation mode. The user can simply ask ChatGPT to translate speech, and the model will continue to translate until it receives a command to stop or switch to another language. This makes the voice mode a more versatile tool in communication.

The feature is available on all devices and platforms, but only for users with a paid subscription. OpenAI did not limit the update by region – voice capabilities are enhanced globally. The new version retains support for dialog communication format and fast response to voice requests.

At the same time, the company warns that minor technical deviations may be observed in the updated mode. Possible problems include unexpected changes in intonation, pitch, and occasional playback failures. Some voice mode “hallucinations”, including meaningless phrases and background music, have not yet been resolved.

ChatGPT’s conversational voice mode was launched in the fall of 2023 and has been actively developed since then. The new version, OpenAI claims, brings the model closer to a more lively and contextual user interaction. The improvements are especially noticeable for longer conversations and translation work.

OpenAI continues to adapt voice functionality for broader tasks. In the future, the company plans to introduce new voice models and improve the response to complex emotions in speech.