OpenAI launches GPT-4o and more features for ChatGPT

If you’re using the free version of ChatGPT, you’re about to get a boost. On Monday, OpenAI debuted a new flagship model of the underlying engine, called GPT-4o, along with significant user interface changes.

The chatbot, which launched a whole new wave of consumer-friendly AI, comes in two flavors: the free version, ChatGPT 3.5, and a version that costs $20 per month, ChatGPT 4. With that subscription fee you get access to a large language model that can process much more data as it generates responses to your prompts.

Look at this: Everything OpenAI announced at the Spring Update event

GPT-4o should close that gap, at least somewhat. Your interactions with ChatGPT will also become more conversational. (The “o” stands for “omni.”)

AI Atlas art badge AI Atlas art badge

When ChatGPT debuted in November 2022, it set off a wave of generative AI products and experiments that continue today. Startups like Anthropic, as well as tech giants like Google and Microsoft, introduced their own spin on the AI ​​chatbot generation, and companies like Rabbit and Humane have built new gadgets around the technology. (See our reviews of these products, as well as advice and news, on our new AI Atlas hub.)

OpenAI’s GPT-4o news came just a day before Google hosts its annual I/O developers conference, where artificial intelligence will likely be a major theme related to its Gemini chatbot and its Search Genative Experience. And in early June, Apple is expected to have a lot to say about AI at its own developer event, WWDC.

Despite ChatGPT 3.5’s limitations, CNET reviewer Imad Khan noted earlier this year that “for most basic questions and even a few more complex ones, ChatGPT 3.5 will get the job done for most people,” providing “usable” answers even though it “It’s always best to do a little fact checking.” Is ChatGPT 4 worth paying more for? In his review of ChatGPT 4, Khan says it’s “noticeably smarter than its free counterpart. And for those who strive for accuracy and ask questions that require more math skills, it’s a worthy upgrade.”

The new GPT-4o model

In a livestream on Monday, OpenAI chief technology officer Mira Murati said GPT-4o “brings GPT-4 level intelligence to everything, including our free users.” The features will be rolled out in the coming weeks, she said.

Paid users get five times the capacity limit of free users.

GPT-4o is two times faster and 50% cheaper than GPT-4 Turbo, which was released in late 2023. (It provided more timely answers than its predecessors and can understand (and generate) larger chunks of text.)

It will be available in 50 languages ​​and is also coming to the API so developers can start building with it.

Talk to the new user interface

Earlier this year, OpenAI removed the requirement to sign up for accounts. On Monday it announced a desktop app, which also aims to make AI tools available to more users.

“We want you to be able to use it wherever you are,” Murati said. “It’s easy, it’s simple, it can be integrated into your workflow very, very easily.”

In another step toward making AI more accessible, OpenAI has announced a “revamped” user interface that includes the ability to interact with ChatGPT on a more conversational level, as well as video sharing as a starting point.

OpenAI CTO Mira Murati takes the stage for the Spring 2024 Update livestream. OpenAI CTO Mira Murati takes the stage for the Spring 2024 Update livestream.

OpenAI CTO Mira Murati on stage for the company’s Spring 2024 Update livestream.

OpenAI/screenshot by CNET

“We know that as these models become more and more complex, we want the experience of interaction to become more natural,” Murati said. “This is the first time that we are really making a huge step forward when it comes to ease of use.”

With the new real-time conversational speech functionality, you can interrupt the model, you don’t have to wait for a response and the model picks up on your emotions, says Mark Chen, head of boundary research at OpenAI.

You can invoke ChatGPT with a familiar phrase: “Hey, ChatGPT.”

Use video and screenshots

Other ways to interact with ChatGPT now include video, so you can share live footage of, say, a math problem you’re stuck on and ask for help solving it. ChatGPT gives you the answer – or helps you work through it yourself.

You can also share screenshots, photos, and documents with text and images, ask ChatGPT about past conversations, search for real-time information within a conversation, and perform advanced data analysis by uploading charts or code before asking questions.

On Friday, OpenAI announced additional data analysis features, including the ability to upload files from Google Drive and Microsoft OneDrive, as well as interact with tables and charts in a new extensible view and customize and download charts.

These features will be available to ChatGPT Plus, Team and Enterprise users “in the coming weeks,” according to a blog post.

Prior to this update, GPT-4, which released in March 2023, was available through the ChatGPT Plus subscription for $20 per month. It uses 1 trillion parameters, or pieces of information, to process queries. An even older version, GPT-3.5, was available for free with a smaller context window of 175 billion parameters.

“We also care deeply about the next frontier,” Murati said. “So soon we will update you on our progress towards the next big thing.”

Look at this: How to create a GPT in ChatGPT-4

Editor’s note: CNET used an AI engine to create several dozen stories, which are labeled accordingly. The note you are reading is linked to articles that delve substantively into the topic of AI, but are created entirely by our expert editors and writers. For more, see our AI policy.

Leave a Reply

Your email address will not be published. Required fields are marked *