Choose Live
For tutoring, brainstorming, language practice, natural interruption, and desktop Work or Codex task coordination.
Mode comparison
GPT-Live vs Advanced Voice vs Standard Voice
Choose Live for the newest natural full-duplex conversation, Advanced for eligible mobile video or screen sharing and current custom-GPT compatibility, and Standard for a simpler transcribe-then-answer flow.
Last verified:
ChatGPT Voice mode matrix
| Capability | Live | Advanced | Standard |
|---|---|---|---|
| Conversation style | Full-duplex; can listen while speaking | Previous realtime experience | Turn-by-turn transcription before response |
| Text and images in the same chat | Yes, when available | Account-dependent | Chat and account-dependent |
| Video and screen sharing | No at launch | Yes for eligible iOS/Android subscribers | No |
| Custom GPTs | Not initially | Voice remains available; uses Shimmer | Account-dependent |
| Web search and memory | Supported when available | Feature/account-dependent | Feature/account-dependent |
| Best reason to choose | Natural interruptions and current Voice experience | Visual mobile sharing or GPT compatibility | Predictable turn-by-turn speech |
Which mode should you use?
Choose Advanced
When you need eligible mobile video, screen sharing, or Voice inside a custom GPT while Live lacks those capabilities.
Choose Standard
When you prefer a clear speak-transcribe-answer cycle or Live and Advanced are unavailable for your account.
The modes can coexist
Live replacing your default experience does not make Advanced or Standard obsolete. Open Settings → Voice to switch among the options exposed to your plan, region, workspace, and app version.
If a Live feature is missing, choose the mode that already supports the task instead of assuming an account bug. For example, video and screen sharing remain an Advanced capability at launch.
Voice mode FAQ
Does Live support video?
No at launch. Eligible iOS and Android subscribers can use video or screen sharing in Advanced Voice.
Can Live run inside a custom GPT?
Not initially. OpenAI says Voice conversations with GPTs remain available through Advanced Voice and use the Shimmer voice.
Why would I use Standard Voice?
Standard transcribes speech before generating a response. That turn-by-turn flow may be preferable when you want a simpler interaction or other modes are unavailable.