For questions that require web search, deeper reasoning, or more complex work, GPT-Live can delegate to our latest frontier model behind the scenes, and brings the result back into the conversation when it's ready.

A good voice model should be enjoyable to talk to, and GPT-Live is a great conversationalist with a more natural and defined personality than previous voice models.

Our full-duplex architecture allows the model to engage in more natural back-and-forth, maintain a better sense of time, and even perform live translation.

GPT-Live is built on a full-duplex architecture, meaning it can listen and speak at the same time.

GPT-Live makes talking with AI feel like having a real conversation. It’s also our smartest voice model yet.

Introducing GPT-Live, a new generation of voice models for natural human-AI interaction. Rolling out in ChatGPT starting today. You’ll want to turn the sound on for this one.

These new models are rolling out to everyone over the next few days, starting today. They’ll be available in ChatGPT across iOS, Android, and web. Coming soon to the API. Just tap the Voice button to talk with ChatGPT, and you’ll feel the magic of GPT-Live.

What kind of deep research would you delegate first? Files, docs, or something more complex? Genuinely curious how others plan to use this.

@ChiaraCervetta Sounds like samanta 😳

Yes, this is where I want routing. And what about memory? I want personalised answers

how do i turn on “show me” ? (this is the new live model)

Does it mean it can now use connectors?

Awesome ‼️

YES PLEASE Now……………..

@nickbaumann_ Man this is my hope for Siri with access to agents and tools for 3rd party app actions

so fucking smart! Nice!

Curious how it signals the behind-the-scenes handoff when it takes longer

Wow

just make it do the thing the people want: control codex

Hello

This feels like a major shift. AI is no longer just responding in real time. It’s actively researching, reasoning, and returning with better answers while the conversation continues. 🎙️🤖

Interesting design choice. Once voice UX delegates to a deeper model in the background, I'd want the full request trail exposed: which route fired, why it switched, and what came back. That's the difference between "felt smart" and production-debuggable.

Live delegation is the workflow layer that changes AI from a search tool into a working partner. I've been framing this for Gen X professionals as the handoff skill — knowing what to keep vs. what to delegate is the new leverage point.

so its basically speed dialing the smart sibling when stuff gets hard

kind of comforting that even the AI knows when to phone a friend for the hard stuff 😄

@jxnlco Is anyone having the issue where gpt live doesn’t have access to long term memory?

That's the key differentiator — routing to the frontier model transparently instead of hiding the handoff. Makes GPT-Live actually useful for real research.

When open source

lady gaga is relying？

Thank you. This was really limiting before.

This delegation feature for GPT-Live sounds like a significant leap in handling complex queries. How does the system decide when to delegate and to which specific model?