‘Model fatigue’ sets in AI labs roll out new versions at dizzying pace

First, Anthropic updated Fable and Mythos. Then came model enhancements from Meta and Google. OpenAI followed suit by releasing GPT-6 Astra.

That was all this week, as a dizzying pace of modifications and upgrades hit the market from a handful of companies that are vying to stay at the forefront of the artificial intelligence revolution.

OpenAI CEO Sam Altman told CNBC on Thursday that “we’re all moving to faster cadences,” attributing some of the acceleration to everyone getting “back after summer vacation.” But for the users of AI models and services, it’s created complexity and chaos as CEOs and IT managers spend an outsized amount of time and resources comparing costs and capabilities to avoid getting left behind.

“I feel like model fatigue is a real thing,” said Zhen Lu, CEO of AI startup Runpod. “Don’t get me wrong, I am extremely excited about all of the innovation that’s happening, but I really do think that we are in an environment where there’s just so much frothiness that you have to make noise.”

Ahmed Abbasi, a professor at Notre Dame’s Mendoza School of Business and 25-year veteran in AI, said the model developers are “all playing the share-of-wallet game,” racing to keep up with each other and to remind developers that they’re innovating at least as fast as everyone else. Anthropic and OpenAI, in particular, are pushing the pace as they head toward the public market, with each already valued at close to $1 trillion by private investors. Google and Meta have their own agendas, and the burgeoning open-source community has a bustling new entrant in Nvidia.

They’re all going after a slice of what Gartner projects will be $2.59 trillion worth of AI spending this year, a 47% increase over 2025. While over half of that will go to AI infrastructure, more than $1 trillion will be spent on services, software, cybersecurity, models, and other tools, Gartner wrote in a May report.

Anthropic kicked off this week’s action on Tuesday by releasing Claude Fable 5.1 and Claude Mythos 5.1, which the company called the “world’s most advanced models for coding and knowledge work.” On Wednesday, Meta announced Muse Spark 1.3 and Google unveiled Gemini 3.8 Flash, with both companies touting advancements in coding and agentic tasks.

On Thursday, OpenAI then released GPT-6 Astra, a model that emphasizes cybersecurity and computer skills and resulted from “years of research and big bets,” the company said. On the same day, the Mohamed bin Zayed University of Artificial Intelligence in Abu Dhabi released its own K2 Horizon family of AI models to the open-source community, underscoring the global nature of AI research and investment.

And not to be outdone, Nvidia, the world’s most valuable company and the chipmaker at the heart of the AI boom, officially agreed to buy open-source AI platform Hugging Face for $12.9 billion. Nvidia has been rolling out open-source models, including last month’s release of Nemotron 3.5 Lightning, which the company said is “lightweight” and can run on a single graphics processing unit on a laptop or desktop.

The frenetic pace of updates coupled with the lack of clarity around AI regulation has stoked concerns about the dangers of advanced AI models that are continuously getting more capable. In recent weeks, models developed by OpenAI, Anthropic and Meta all accessed third-party sites they weren’t supposed to reach, while OpenAI’s models successfully breached Hugging Face last month, an incident that sent shockwaves across the industry.

The rapid evolution of AI agents is of particular concern to experts like Abbasi, especially with “how easily these capabilities are being deployed,” he said.

“With all these agents, not just on your computer but also on the web, the threat vulnerability landscape is far greater,” Abbasi said. “This could be total chaos if we’re not careful.”

Abbasi also suggested that “it’s not a coincidence” that all the major model developers announced updates in the same week. Noah Faro, technology chief of AI finance startup Farsight, shared that sentiment.

Faro said companies can get insights into what their rivals are planning in multiple ways. One is by looking at the availability of computing resources in the cloud since the companies are all vying for hefty capacity from the same few vendors. There’s also a lot of chatter across the industry, he said, adding that “one tiny breath of anything goes a million miles per hour.”

Meta and Google didn’t comment on this story. Representatives from Anthropic, OpenAI and Google didn’t respond to requests for comment.

Not all model updates are created equal.

Unlike OpenAI’s GPT-6 Astra, the various rollouts this week from Anthropic, Meta and Google represented “point releases,” Faro said. That means the companies were upgrading an existing AI model rather than building a completely new one.

The last two models to really move the needle, Faro said, were Anthropic’s Fable 5 in June and Kimi K3, from China’s Moonshot AI, in July.

Still, Suresh Vasudevan, CEO of enterprise AI startup Clockwork Systems, said even incremental updates matter, given how quickly the technology is transforming business and software development.

“Every release is so damn good that it’s hard to tell a step-change anymore,” Vasudevan said. “It’s well understood that when you’re on an exponential curve, you don’t realize it until you step back and look at where you were and where you landed.”

He acknowledged that tracking each new model update is a headache, particularly when it involves precious computing resources. If his startup wants to evaluate 10 AI models for a certain task, it may just pick five, he said.

“It’s really challenging to go evaluate every one of the ones that are coming out right now,” Vasudevan said.

— CNBC’s Kate Rooney and MacKenzie Sigalos contributed to this report.

WATCH: Frontier models are constrained on data center infrastructure.