OpenAI's "Spring Update" revealed new updates to ChatGPT.
OpenAI CTO Mira Murati kicked off the event.
She announced GPT-4o, its next flagship AI model, with improved voice and vision capabilities.
OpenAI just took the wraps off a big new update to ChatGPT.
Cofounder and CEO Sam Altman had teased "new stuff" coming to ChatGPT and GPT-4, the AI model that powers its chatbot, and told his followers to tune in Monday at 1 p.m. ET for its "Spring Update" to learn more.
Also ahead of time, Altman ruled that the event would reveal GPT-5 or a new OpenAI search engine, which is reportedly in the works. OpenAI is reportedly planning to eventually take on internet search giant Google with its own AI-powered search product.
But the big news on Monday was OpenAI's new flagship AI model, GPT-4o, which will be free to all users and "can reason across audio, vision, and text in real time." It was CTO Mira Murati who delivered the updates with no appearance on the livestream from Altman.
There were a ton of demos intended to demonstrate the real-time smarts of GPT-4o.
OpenAI researchers showed how the new ChatGPT can quickly translate speech and help with basic linear algebra using its visual capabilities. The use of the tech on school assignments has been a polarizing topic in education since it first launched.
Say hello to GPT-4o, our new flagship model which can reason across audio, vision, and text in real time: https://t.co/MYHZB79UqN
Text and image input rolling out today in API and ChatGPT with voice and video in the coming weeks. pic.twitter.com/uuthKZyzYx
OpenAI posted another example to X of how one can interact with the new ChatGPT bot. It resembled a video call, and it got pretty meta.
In the video, ChatGPT takes in the room around it, discerns it's a recording setup, figures it might have something to do with OpenAI since the user is wearing a hoodie, and then gets told that the announcement has to do with the AI — it is the AI. It reacts with a voice that sounds more emotive.
OpenAI also announced the desktop version of ChatGPT, and a new and improved user interface.
In addition to GPT-4o and ChatGPT, OpenAI's other products include its AI-powered image generator DALL-E, its unreleased text-to-video generator Sora, and its GPT app store.
You can catch up on our liveblog of the event below.
That’s a wrap! OpenAI concludes the event without an appearance from Altman.
OpenAI says text and image input for GPT-4o-powered ChatGPT is launching today. Meanwhile, voice and video options will drop in the coming weeks, the company said.
Although Altman didn't step in front of the camera, the CEO posted videos from the audience on X.
He also teases "more stuff to share soon."
GPT-4o can also break down charts
The new AI model can interact with code bases, the OpenAI execs say. The next demo shows it analyzing a chart from some data.
It's a plot of global temperatures. GPT-4o gives some takeaways from what it sees, and CTO Mira Murati asks about the Y axis, which the AI explains.
ChatGPT reads human emotions — with a stumble
OpenAI
For the last live demo of the day, Zoph holds his phone up to his face and asks ChatGPT to tell him how he looks. Initially, it identifies him as a "wooden surface" — a reference to an earlier photo he had shared.
But after a second try, the model gives a better answer.
"It looks like you're feeling pretty happy and cheerful," ChatGPT says, noting the small smile on Zoph's face.
In one of the final tests, ChatGPT becomes a translator
OpenAI
In response to a request from an X user, Murati speaks to ChatGPT in Italian.
In turn, the bot translates her query into English for Zoph and Chen.
"Mike, she wonders if whales could talk, what would they tell us?" she said in English after hearing Murati's Italian.
It's pretty impressive.
The video demo shows how it could help with math homework, including basic linear algebra
OpenAI
OpenAI Research Lead Barret Zoph walks through an equation on a whiteboard (3x+1=4), and ChatGPT gives him hints as he finds the value of x — making it basically a real-time math tutor.
At the beginning, the bot jumped the gun.
"Whoops, I got too excited," it said after it tried to solve the math problem hadn't been uploaded yet.
But it then walked him through each step, recognizing his written work as he tried to solve the equation.
It was able to recognize math symbols, and even a heart.
OpenAI's first demo: Talking to GPT-4o
It's demo time!
The new bot has a voice that sounds like an American female, but no word yet if you can change it.
OpenAI Research Lead Mark Chen pulled out ChatGPT on his phone and asks for advice on giving a live presentation using Voice Mode.
"Mark, you're not a vacuum cleaner," it responds when he hyperventilates, appearing to perceive his nervousness. It then tells him to moderate his breathing.
Some big changes, you can interrupt the AI now, and there shouldn't be the usual 2 or 3-second delay with GPT-4o.
It can also detect emotion, according to OpenAI.
GPT-4o will have improved voice capabilities
OpenAI
Murati emphasizes the necessity of safety with the real-time voice and audio capabilities of the new GPT-4o model.
She says OpenAI is "continuing our iterative deployment to bring all the capabilities to you."
Murati says the big news is a "new flagship model" called GPT-4o.
The new model is called GPT-4o, and Murati says that OpenAI is making a "huge step forward" with ease of use with the new model.
It's free for users, and "allows us to bring GPT-4 class intelligence to our free users," Murati says.
And we're off!
OpenAI CTO Mira Murati
OpenAI
The livestream began with CTO Mira Murati at OpenAI's offices.
OpenAI is going to be announcing 3 things today, she says. "That's it."
For those who want to watch live, you can view the whole event here.
OpenAI will be livestreaming its spring update, which kicks off in less than an hour.
