ChatGPT Voice moves to GPT-Live-1.1 and adds new voices for six languages
ChatGPT Voice is getting a new model and a wider choice of voices. OpenAI says it is rolling out GPT-Live-1.1 and a smaller version called GPT-Live-1.1 Quip, with more natural accents across languages.
OpenAI's example of a weather card that ChatGPT Voice can show while it talks. Image: OpenAI
New voices cover French, German, Hindi, Japanese, Korean and Mexican Spanish, plus a new English voice called Bubbie. OpenAI says the voices each person sees may vary by region and account.
Six languages get voices of their own, with more natural accents
The accent work is the main change for anyone who talks to ChatGPT in a language other than English. The new voices give French, German, Hindi, Japanese, Korean and Mexican Spanish speakers options built for their language.
To pick one, open Settings, then Voice, and choose Voice.
Which model you hear depends on your plan. OpenAI's Voice help page lists GPT-Live-1.1 for Plus, Pro and Business users, and the smaller Quip for Free and Go. Plus users get three hours a day with the full model, while Pro at $200 per month has no fixed cap.
OpenAI's example of a sports card from ChatGPT Voice. Image: OpenAI
ChatGPT Voice now tells you more clearly when it is searching the web
The update also changes how Voice fills the pause. OpenAI says GPT-Live now uses a wider range of short phrases to let you know when it is searching. That spoken cue helps most when you cannot see the screen.
GPT-Live is the model family behind ChatGPT Voice. It can also put visual answers such as weather, stock and sports cards on screen while it talks.
OpenAI's example of a map answer from ChatGPT Voice. Image: OpenAI
Voice is how smart glasses owners reach ChatGPT when there is no screen
For anyone using ChatGPT away from a keyboard, Voice is the interface. Smart glasses owners already reach ChatGPT through apps such as OpenVision, which runs Ray-Ban Meta glasses on a ChatGPT subscription. A more natural accent in German or Japanese makes that kind of use less awkward for people outside the US.
The rollout is gradual, so not every account will see the new voices yet.
