Microsoft-backed OpenAI announces GPT-4 Turbo, its most powerful AI yet

OpenAI announced its new, more powerful GPT-4 Turbo artificial intelligence model in the US overnight Monday during its first in-person event, and revealed a new option that will let users create custom versions of its viral ChatGPT chatbot.

It’s also cutting prices on the fees that companies and developers pay to run its software.

OpenAI’s announcements show that one of the hottest companies in tech is rapidly evolving its offerings in an effort to stay ahead of rivals like Anthropic, Google and Meta in the AI arms race.

ChatGPT, which broke records as the fastest-growing consumer app in history months after its launch, now has about 100 million weekly active users, OpenAI said Monday. More than 92 per cent of Fortune 500 companies use the platform, up from 80 per cent in August, and they span across industries like financial services, legal applications and education, OpenAI CTO Mira Murati told reporters Monday.

The event also included a surprise appearance by Microsoft CEO Satya Nadella.

“The systems that are needed as you aggressively push forward on your road map require us to be on the top of our game, and we intend fully to commit ourselves fully to making sure you all... have not only the best systems for training and inference, but also the most compute,” Nadella told OpenAI CEO Sam Altman while onstage together.

He added, “That’s the way we’re going to make progress.”

Earlier this year, Microsoft’s expanded investment in OpenAI — an additional $US10 billion ($15.4b) — made it the biggest AI investment of the year, according to PitchBook, and in April, the startup reportedly closed a $US300 million share sale at a valuation between $US27b and $US29b, with investments from firms such as Sequoia Capital and Andreessen Horowitz.

As recently as last month, OpenAI was reportedly in talks to close a deal that would lead to an $US80b valuation.

In his speech Monday, Altman said the day’s announcements came from conversations with developers about their needs over the past year. And when it comes to GPT-5, Altman told reporters, “We want to do it, but we don’t have a timeline.”

Here’s what OpenAI announced Monday:

GPT-4 Turbo

GPT-4 Turbo is the latest AI model, and it now provides answers with context up to April 2023. Prior versions were cut off at January 2022. For example, if you asked GPT-4 who won the Super Bowl in February 2022, it wouldn’t have been able to tell you. GPT-4 Turbo can.

“We are just as annoyed as all of you, probably more, that GPT’s knowledge about the world ended in 2021,” Altman said in a speech Monday.

It also accepts a lot more input. While earlier versions limited you to about 3000 words, the GPT-4 Turbo accepts inputs of up to 300 pages in length. It means you could ask it to summarise entire books.

GPT-4 also supports DALL-E 3 AI-generated images and text-to-speech. It also has six preset voices to choose from, so you can choose to hear the answer to a query in a variety of different voices.

OpenAI said GPT-4 Turbo is available in preview for developers now and will be released to all in the coming weeks.

OpenAI said it’s also cutting the prices for developers. “GPT-4 Turbo input tokens are 3x cheaper than GPT-4 at $US0.01 and output tokens are 2x cheaper at $US0.03,” the company said, which means companies and developers should save more when running lots of information through the AI models.

Personalised chatbots

Until now, ChatGPT’s enterprise and business offerings were the only way people could upload their own data to train and customise the chatbot for particular industries and use cases. Now it’s adding the option for anyone to create custom chatbots.

AI “agents” are one of the buzziest uses of the technology recently, with many startups vying to offer the kind of personalised AI tools that consumers may already be familiar with via pop culture representations, such as Tony Stark’s J.A.R.V.I.S. in Marvel movies, or Pam in Disney Channel’s Smart House.

“Anyone can easily build their own GPT — no coding is required,” the company wrote in a release. “You can make them for yourself, just for your company’s internal use, or for everyone. Creating one is as easy as starting a conversation, giving it instructions and extra knowledge, and picking what it can do, like searching the web, making images or analysing data.”

More than two million developers building their own tools using ChatGPT’s API will also be able to customise the chatbot, meaning consumers will likely see personalized AI chatbots popping up in many more places, including apps and websites they use regularly.

Open AI’s version of the App Store

Now that users and developers can launch their own, personalised AI chatbots, OpenAI is introducing a new revenue driver for the company: Its own version of the app store.

The GPT Store allows people who create their own GPTs to make them available for public download, and in the coming months, OpenAI said people will be able to earn money based on their creation’s usage numbers.

“Once in the store, GPTs become searchable and may climb the leaderboards,” the company wrote in a release. “We will also spotlight the most useful and delightful GPTs we come across in categories like productivity, education, and ‘just for fun.’”

As for revenue share for people who create custom chatbots featured in the store, the company will start with “just sharing a part of the subscription revenue overall,” Altman told reporters Monday. Right now, the company is planning to base the payout on active users plus category bonuses, and may support subscriptions for specific GPTs later.

“What OpenAI is really in the business of selling is intelligence -- and that, and intelligent agents, is really where it will trend over time,” Altman told reporters.

New all-in-one image-generation, browsing and summarisation

Until Monday, ChatGPT users had to hop between different apps and websites to use all of OpenAI’s tools, which contributed to a slightly higher learning curve.

On Monday, the company announced it has streamlined its AI tools into one place: Using ChatGPT now offers image generation via DALL-E, browsing, data analysis, document upload and PDF search. Before now, Anthropic’s Claude was the only competitor chatbot to allow PDF search.

Copyright shield

As generative AI-related legal action is on the rise, Altman announced Monday that OpenAI will “step in and defend our customers” and “pay the costs incurred if you face legal claims around copyright infringement.” It echoes similar statements made by Google, Microsoft and Adobe.

CNBC