New GPT-4o voice mode features are coming out

Key Takeaways

• GPT-4o Voice Mode will improve the natural experience of talking with ChatGPT.
• New features include reduced response time and different tones of voice.
• An initial release to a select group of ChatGPT Plus subscribers, with a wider release expected in the fall.

After waiting longer than expected, OpenAI's Sam Altman revealed in his response to X that the new GPT-4o voice features will finally start rolling out next week. However, this alpha release will be limited to a small set of ChatGPT Plus subscribers at first, features that will likely see a wider release sometime in the fall.

Back in May, OpenAI showed off the GPT-4o, a new model. The show included impressive new capabilities, such as the ability to respond to information from a real-time video feed, and new voice features that would make talking to the GPT-4o seem like talking to a human. When the GPT-4o was released, voice capabilities were missing, with messages in the app indicating that new Voice Mode features would be released soon. Now it seems that the release will finally begin.

GPT-4o Voice will make talking to ChatGPT feel more natural

The voice will be able to work better and will have more abilities

Even before the introduction of GPT-4o, you already talk to GPT-4 in Voice Mode, but one of the biggest problems is that it is difficult to have what sounds like a natural conversation when there is a delay of 5.4 seconds. You speak out loud, and then you have to watch the thought bubble animation for a few seconds before you get any response.

The new Voice Mode of the GPT-4o will reduce the average response time down to 320 milliseconds and can go as low as 232 milliseconds. This allows you to have what sounds like a quick back-and-forth conversation with the GPT-4o. At the shows at the time of the announcement, the responses were surprisingly quick. It is also possible to interrupt the response by speaking again; the voice response will stop and the GPT-4o will start listening again.

If the field skills are as impressive as they are on display, it will make talking to the GPT-4o almost like talking to someone else.

However, speed is not the only change. It is possible to get the GPT-4o to speak in different tones or in different ways. Demonstration videos show the GPT-4o speaking in a sarcastic voice, speaking like a sportscaster, counting to ten at different speeds, and singing happy birthday. If the field skills are as impressive as they are on display, it will make talking to the GPT-4o feel like talking to another person.

The Voice Mode on the GPT-4o is also capable of real-time translation. For example, it is possible for one person to speak to the GPT-4o in one language and a second person to speak to the GPT-4o in a different language. The GPT-4o will then repeat each phrase in the opposite language, allowing two people who do not speak the same language to hold a conversation.

You will have to wait for GPT-4o voice mode for a while

New features are released to a small group of ChatGPT Plus users

The first release of new features has been a long time coming. OpenAI said in May that it would be deployed in the “coming weeks” but the number of weeks since the announcement has doubled. However, the wait is almost over, at least for a small number of people. Along with an endorsement from Sam Altman at X, a message inside the ChatGPT app also says that Open AI “will begin alpha with a small group of Plus users in late July.”

This small initial release means that even if you're a ChatGPT Plus user, it's very unlikely that you'll get access to the new Voice Mode features next week. However, the message also says “the plan is for all Plus users to have access in the fall” so hopefully, we won't all have too long to wait. One thing is certain; when the new Voice Mode drops, it won't sound like Scarlett Johansson.