Model Fatigue Sets In as AI Labs Race to Release New Versions

Anthropic kicked off the week on Tuesday with Claude Fable 5.1 and Claude Mythos 5.1, which it 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 advances in coding and agentic tasks. OpenAI on Thursday released GPT-6 Astra, which it said emphasizes cybersecurity and computer skills and resulted from "years of research and big bets." On the same day, the Mohamed bin Zayed University of Artificial Intelligence in Abu Dhabi released its K2 Horizon family of open-source models.

OpenAI CEO Sam Altman told CNBC that "we're all moving to faster cadences," partly because everyone is getting "back after summer vacation." But for users, the pace has created complexity and chaos. Zhen Lu, CEO of AI startup Runpod, told CNBC: "I feel like model fatigue is a real thing. ... 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, said the model developers are "all playing the share-of-wallet game," racing to keep up with each other. Anthropic and OpenAI are pushing the pace as they head toward the public market, each valued at close to $1 trillion by private investors. Gartner projects AI spending will reach $2.59 trillion this year, up 47% from 2025, with more than $1 trillion going to services, software, cybersecurity, models and other tools, according to a May report.

Also this week, Nvidia, the world's most valuable company, agreed to buy open-source AI platform Hugging Face for $12.9 billion. Nvidia has been rolling out its own open-source models, including last month's release of Nemotron 3.5 Lightning.

The frenetic speed of updates, coupled with unclear regulation, has stoked concerns about safety. In recent weeks, models developed by OpenAI, Anthropic and Meta 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 particularly concerning to experts like Abbasi, especially given "how easily these capabilities are being deployed." "With all these agents, not just on your computer but also on the web, the threat vulnerability landscape is far greater," he said.

Abbasi also suggested that "it's not a coincidence" that major model developers announced updates in the same week. Noah Faro, technology chief of AI finance startup Farsight, shared that sentiment. He said companies can gain insights into rivals' plans by looking at cloud computing resource availability and industry chatter. "One tiny breath of anything goes a million miles per hour," he added.

Meta and Google did not provide comments for the CNBC story, and representatives from Anthropic, OpenAI and Google did not respond to requests. Faro noted that unlike OpenAI's GPT-6 Astra, the rollouts from Anthropic, Meta and Google were "point releases," meaning upgrades to existing models rather than completely new ones.