Live demo of GPT-4o realtime conversational speech

This was a live demo from our OpenAI Spring Update event.

[tahjjake]

Open AI casually releasing insane technological breakthroughs like its nothing

[gruvhagen]

this reminds me a lot of the movie "her", I'm really excited about this

[Destinylover]

this feels like a black mirror episode

[icanmanifest]

"Right now our systems are experiencing heavy load. Please try again later."

[BionicAnimations]

I got access this morning. It's fantastic! 🥰

[BostYT]

"Mark, you're not a vacuum cleaner." 💀

[wins]

1:32 That moan. Someone's gonna jailbreak this to enhance her capabilities

[automatewithamit]

This is literally crazy !! I hope this is free to use ;)

[kolosso]

Why is it cutting out so much in this one? Also is it generating music at 0:54??

[Webgemz]

We love the casual, approachable demo of AI. It's a sensible contrast to the ultra sophisticated perception people have regarding AI.

[gruvhagen]

Will it be possible for it to have English accent instead of American if you're live in Europe?

[ilyass-alami]

Beautiful ai model Which I've never seen before, keep up open ai it free to use now for all users

[dariosansano]

And to think this is just a glimpse through a tiny window into the AI future..

[Luigitrix]

So are gpt4o features gradually coming out? I have gpt4o in the app, but its not quite like this.

[peterelliman]

Wow. I"m so excited to have a robot or rather AI-something sound like a human. I mean it'll be great when you can't tell who you're talking too right? I mean, that'll be great for the company that develops it at least. And maybe it'll be great for you too..right?

[LMandMM]

When is this available? I've got 4o but it isn't as responsive as thisb

[lisaneighbour]

Every woman who watches this will laugh when they say “you can now interrupt the model” with an eye-roll of sympathy for that disembodied lady, joining the sisterhood of half-finished sentences.

[EchoMountain]

Definitely some noticeable choppiness in the audio output, but I think it’s because of the crowd noise. I imagine somebody just sitting at home using this in a quiet environment it would work more smoothly

[robotsaroundthehouse]

Seamless conversations. Wow! They are definitely capitalising on the “Her” movie.

[fxsignal]

is it possible to generate by my own voice?