OpenAI has launched advanced voice mode for audio chats with ChatGPT.

Jordan Novet, CNBC

ChatGPT is ready for more natural audio chats.

OpenAI said Tuesday that its popular chatbot now has an advanced voice feature for people who pay for the premium service. The tool allows for more fluid conversations.

The release will continue through the week. The company said it’s not yet available in EU countries, Iceland, Liechtenstein, Norway, Switzerland or the U.K.

OpenAI announced the new capability in May. The rollout got plenty of publicity because of a voice called Sky that resembled that of Scarlett Johansson in the 2013 movie “Her.” Legal counsel working on behalf of Johansson sent OpenAI letters claiming the company didn’t have the right to use the near-identical voice, and OpenAI paused using it in its products, CNBC reported.

In the months since, people have been able to configure ChatGPT to talk to them in other voices through a free tier. The advanced version responds more quickly and will stop talking and listen if you interrupt it. There are now nine voices to choose from, and you can enter instructions for voice chats in the Customizations part of the app’s settings.

“Hope you think it was worth the wait,” Sam Altman, OpenAI’s co-founder and CEO, wrote in an X post on Tuesday.

It’s an increasingly competitive space for OpenAI, which is backed by Microsoft.

For the past couple weeks, Google has been releasing its own Gemini Live voice feature in English on Android devices. And on Monday, Reuters reported that Meta will introduce celebrity voices later this week that can be accessed through Facebook, Instagram and WhatsApp.

OpenAI got a head start in the generative AI chatbot market, when it launched ChatGPT in late 2022. In August, OpenAI told media outlets that ChatGPT had over 200 million weekly active users.

The advanced mode is only available to those with subscriptions to OpenAI’s Plus, Team or Enterprise plans. The most affordable option is the Plus tier at $20 per month.

Here’s what to do

If you’re paying, it’s easy to get started, assuming OpenAI has granted access to your device.

First, make sure you have the latest version of the app on your phone.

Open the ChatGPT app.

OpenAI says you’ll receive a notification in the app once it’s turned on access to the new feature. Hit the continue button to get started.

Create a new chat by swiping right or tapping the two-line icon in the top left corner and selecting ChatGPT at the top. To the right of the “Message” text field and the microphone icon, you should see a sound wave icon. Tap that and make sure your sound is on.

In a few seconds, you’ll hear a little “bump” sound, and the circle in the middle of the screen will morph into a fluid sky-like blue and white animation. Start speaking. You should receive a response quickly. Don’t be surprised if audio breaks up a bit.

OpenAI said it has improved accents in some foreign languages and has increased the speed of conversations. But if you don’t like what you’re hearing, you can ask ChatGPT to speak differently. You can tell it to speed up, for example, or incorporate a Southern accent.

With advanced voice mode, you can have ChatGPT tell you a bedtime story, help you get ready for a job interview or even work on your foreign language skills.

But even if you’re paying, you won’t have unlimited access to advanced voice mode. After using it for about a half hour on Tuesday, I started seeing “15 minutes left” at the bottom of the screen.

OpenAI didn’t immediately respond to a request for details about the time limit.