OpenAI rolled out GPT-Live-1.1 and a smaller GPT-Live-1.1 Quip to ChatGPT Voice on October 15, 2026, along with new voices in several languages. The same documentation also spells out how many hours of talking each plan gets, which decides whether you will actually notice the upgrade.
If you use ChatGPT Voice on a commute or while cooking, the model behind the call now depends on what you pay for.
TL;DR: ChatGPT Voice now runs on GPT-Live-1.1 for Plus and higher plans and on the lighter Quip model for Free and Go. The update adds voices for French, German, Hindi, Japanese, Korean and Mexican Spanish, plus an English voice called Bubbie. Talk time ranges from limited on Free to three hours on Plus and unlimited on the $200 Pro plan. Video and screen sharing still need the older Advanced mode.
What the GPT-Live-1.1 update changes in ChatGPT Voice
According to OpenAI’s release notes, GPT-Live-1.1 and GPT-Live-1.1 Quip are rolling out in ChatGPT Voice, and Quip is the mini version. The update brings more natural accents across languages and a wider range of short phrases that signal when Live is searching.
That second change is small but practical, and it builds on the original GPT-Live rollout. Anyone who has sat through a silent pause wondering whether the assistant froze will see why a spoken cue helps.
The new voices cover French, German, Hindi, Japanese, Korean and Mexican Spanish, plus Bubbie in English. To switch, open Settings, then Voice, then select Voice. OpenAI notes that available voices can vary by region and account, and the model you hear is not the same for everyone either.
How much ChatGPT Voice time each plan gets
The ChatGPT Voice help page ties the model to your plan. Quip appears on Free and Go, while GPT-Live-1.1 appears on Plus and higher. Limits run on a rolling 24 hours in Chat.
| Plan | Model | Limit per rolling 24 hours |
|---|---|---|
| Free | Quip | Limited, may change |
| Go | Quip | 3 hours |
| Plus | GPT-Live-1.1 | 3 hours |
| Pro ($100 per month) | GPT-Live-1.1 | 15 hours |
| Pro ($200 per month) | GPT-Live-1.1 | Unlimited, subject to abuse guardrails |
| Business Standard | Not listed | 3 hours |
| Business Premium | Not listed | 15 hours |
For most people three hours is plenty, so the real gap between Go and Plus is the model, not the clock. Where you can spend those hours matters just as much.
Where ChatGPT Voice Live works and where it does not
Voice in Chat runs on web, iOS and Android, and Voice in Desktop covers macOS and Windows. Live can listen and speak at once, so you can interrupt it, and it supports web search, memory and connected apps.
Video and screen sharing are not part of Live. For those, choose Advanced under Settings, then Voice, and use the camera button on iOS or Android.
Hands-free use has one catch. Plugin actions that need approval must be approved on screen, because spoken approval is not supported. Then there is the question of what happens to the recording itself.
What happens to your ChatGPT Voice audio
OpenAI says audio from Live and Advanced is stored with the chat transcript for 30 days. Deleting the chat removes the clips within 30 days, while archiving it keeps them.
Models are not trained on your audio or video unless you opt in under Settings, then Data Controls. Transcripts are a separate matter, since they may be used for training if Improve the model for everyone is on.
Voice is also different from uploading a finished recording, which paid plans can now do through audio uploads. Those rules shape who this update suits best.
Who will notice the GPT-Live-1.1 upgrade
Plus and Pro users get the full model, so they are the likely audience for the accent and search cue changes. Free and Go users get Quip, which OpenAI describes only as the mini version, and the help page does not document technical differences between the two.
If you talk to ChatGPT in French, German, Hindi, Japanese or Korean, the new voices may be the more visible change. Check the voice picker and the limits on your own plan before you build a routine around it.
