Will GPT-5 let OpenAI break away from the pack?

s·Eye on AIWill GPT-5 let OpenAI break away from the pack?By Sharon GoldmanBy Sharon GoldmanAI ReporterSharon GoldmanAI ReporterSharon Goldman is an AI reporter at Fortune and co- Eye on AI, Fortune’s flagship AI

AI work helpers duke it out

Zoox gets a special exemption

One of the defining truths the world of generative AI is that even when you’re on top, the lead doesn’t last for long

And so, the two key questions coming out of OpenAI’s long-awaited launch of GPT-5 today are whether the new LLM can help the company reclaim the mantle of undisputed AI leader—and if so, how long can OpenAI keep the lead? OpenAI says GPT-5 ders “more accurate answers than any previous reasoning model,” and is “much smarter across the board,” reflected by strong performance on academic and human-evaluated benchmarks

Its re blog boasts of new state-of-the-art performance across math, coding, and health questions, and found that GPT-5 outperformed other OpenAI models across tasks spanning over 40 occupations, including law, logistics, sales, and engineering. “GPT-5 really feels talking to a PhD level expert in any topic,” OpenAI CEO Sam Altman told journalists in a pre-briefing on Wednesday. “Something GPT-5 would be pretty much unimaginable in any other time in history.” Altman described GPT-5 as a “significant step” along the path to artificial general intelligence (AGI), which, according to OpenAI’s mission statement, is defined as “highly autonomous systems that outperform humans at most economically valuable work.” It’s un whether this combination of speed, power and features will be enough, however

In addition to Big competitors Meta and Google, there’s a wave of startups founded by OpenAI’s own former reers, including Anthropic, Thinking Machines, and Safe Superintelligence

And of course, there’s the new crop of powerful Chinese models, DeepSeek, vying for global influence

Whether GPT-5 pels OpenAI back to the top of the AI hill will become in the days and weeks ahead, as reers put the model through its paces, testing it against the s of other elite models, including Anthropic’s Claude model and Google’s Gemini

OpenAI pushes to stay in the lead With GPT-5 now finally out, OpenAI CEO Sam Altman acknowledged that staying at the frontier means one thing: relentless scaling

In AI, scaling refers to the idea that models get more powerful as you increase the amount of data, computing power, and model components used during training

It’s the underlying principle that drove gress from GPT-2 to GPT-3 to GPT-4—and now GPT-5

The catch is that each leap requires exponentially more investment, particularly in AI infrastructure—for OpenAI, that includes its Stargate ject, a joint venture it announced in January with Softbank, Oracle and investment firm MGX with a goal to to invest up to $500 billion by 2029 in AI-specific data centers across the U.S

When asked whether scaling laws still hold, Altman said they “absolutely” do

He pointed to better models, smarter architectures, higher-quality data, and significantly more computing power as the path to “order-of-magnitude” imvements still ahead

But that kind of gress comes at a cost. “It’s going to take an eyewatering amount of compute,” he admitted. “But we intend to continue doing it.” That of course, requires massive amount of money and partnerships

On the bright side, OpenAI has roughly doubled its revenue in the first seven months of 2025, hitting an annualized run rate of $12 billion—up from $6 billion at the start of the year, according to a recent report by The Information

And earlier this week OpenAI released a free, open-source model—an unusual move for a company often criticized for its closed apach over the past half-decade—suggesting confidence that its premium offering, which is now GPT-5, will continue to dominate

But there are some big challenges ahead, however

And the Stargate ject is reliant on partnerships with companies SoftBank, whose investment stipulations are tied to OpenAI’s still unresolved efforts to overhaul its corporate structure

There’s much more to say GPT-5

Journalists received the full set of materials from OpenAI, including a re blog, system card, and safety card, a mere 90 minutes before releasing the model, so I still have much more to go through

Stay tuned for more! Also: In less than a month, I will be headed to Park City, Utah, to participate in our annual Brainstorm conference at the Montage Deer Valley! Space is limited, so if you’re interested in joining, register here

I highly recommend: There’s a fantastic lineup of speakers, including Ashley Kramer, chief revenue officer of OpenAI; John Furner, president and CEO of Walmart U.S.; Tony Xu, founder and CEO of DoorDash; and many, many more! With that, here’s the rest of the AI news

OpenAI has struck a deal with the U.S. government to vide federal agencies access to its frontier AI models—including ChatGPT—for just $1 over the next year, according to a joint announcement with the General Services Administration (GSA)

The partnership reflects months of behind-the-scenes outreach by OpenAI CEO Sam Altman and his team, who have been cultivating ties with the Trump administration since before Donald Trump’s return to the White House in January

In a statement to Wired, Altman framed the agreement as part of President Trump’s AI Action Plan, saying it would help public servants leverage AI to better serve the American people.AI work helpers duke it out

One week after OpenAI released a ChatGPT Study Mode, Google has jumped into the AI tutor game with a new Guided Learning mode

Google says it worked closely with learning experts, teachers, and students to develop Guided Learning, which is integrated within Gemini

ChatGPT’s Study Mode, Google’s Guided Learning is designed to encourage users to work through blems through “bing and open- questions,” rather than just spit out the answers

With classes at many schools and universities set to start in the coming weeks, the big question will be whether students actually want an AI study companion, or are content with an AI answer machine

Zoox gets a steering wheel exemption

The evolution of the robotaxi took a step forward Wednesday as U.S. federal regulators gave permission for Amazon’s Zoox to test self-driving shuttles with no steering wheel, pedals, or other manual controls

The move s up a point of ambiguity in the self-driving car industry

While Zoox had previously claimed its vehicles were compliant with federal motor vehicle standards, the National Highway Traffic Safety Administration’s rules seemed to suggest that was not the case

The NHTSA’s move makes it that U.S.-built self-driving cars do indeed need manual controls to operate (even for testing) on public roads, but grants Zoox an exemption—and it opens the door for other robotaxi companies such as Tesla to get a waiver

Oct. 6-10: World AI Week, Amsterdam Oct. 21-22: TedAI San Francisco

Dec. 2-7: NeurIPS, San Diego Dec. 8-9: Fortune Brainstorm AI San Francisco

The companies are pouring huge sums of money into building and expanding data centers filled with expensive Nvidia GPUs in a race to stay ahead in AI and to achieve AGI

This is the online version of Eye on AI, Fortune's weekly on how AI is shaping the future of . for free.