OpenAI has rolled out an exciting update to Advanced Voice Mode for ChatGPT, bringing noticeable improvements to paid users across all markets and platforms. This enhancement focuses on making conversations feel more natural and engaging, with refined intonation, realistic pacing, and better expression of emotions like empathy and sarcasm. Whether you're chatting casually or navigating a complex discussion, the upgraded voice feature aims to deliver a smoother, more human-like experience.

One of the standout additions is seamless language translation. Now, users can ask ChatGPT to translate between languages during a conversation, this is perfect for real-world scenarios like ordering at a restaurant in Brazil or collaborating with a colleague in Tokyo. For example, speak in English, and Voice Mode will translate your words into Portuguese, then relay the waiter's response back in English, keeping the exchange effortless and intuitive. This feature continues translating until you choose to stop or switch languages.

This update builds on earlier improvements from 2025, which reduced interruptions and enhanced accent accuracy. To access the upgraded Advanced Voice Mode, paid users can simply tap the Voice icon in the ChatGPT message composer.

Limitations

While the update brings clear benefits, OpenAI notes some limitations. In testing, minor audio quality dips, such as slight variations in tone or pitch, have occasionally occurred, particularly with certain voice options. The team is working to improve consistency moving forward. Additionally, rare instances of "hallucinations" in Voice Mode, unintended sounds like gibberish, ads, or background music, may still happen, but OpenAI is actively addressing this.

This upgrade marks a meaningful step forward for ChatGPT’s paid users, offering a more natural and versatile voice experience, especially for multilingual needs.