OpenAI said Friday it is restricting the release of its new GPT-5.6 Sol model at the request of the Trump administration, while rival Anthropic announced hours later that the government had approved a limited release of its strongest cybersecurity model - two weeks after the Commerce Department effectively banned it. Both companies are navigating an unprecedented federal vetting process for AI products that carries direct consequences for cybersecurity professionals, government contractors, and developers who rely on frontier models.
The two San Francisco-based companies said their newest models would be available only to small groups of trusted partners. OpenAI's GPT-5.6 Sol, pronounced "SOHL" like the Spanish word for sun, will be accessible solely to customers approved by the Trump administration. The company said roughly 20 customers have been approved so far but declined to name them.
"We don't believe this kind of government access process should become the long-term default," OpenAI said in a statement. The company described the testing period as a temporary step on the "path to broader availability in the coming weeks."
Anthropic's models caught in the dragnet
Anthropic took offline two new AI models - Fable 5 and Mythos 5 - just days after unveiling them to comply with a Trump directive blocking their use by foreign nationals. The government lifted restrictions on Mythos 5 on Friday, enabling it to be "redeployed to a small group of cyber defenders and infrastructure providers," the company said. Fable 5, which Anthropic had pitched as a safer version of Mythos, remains unavailable even after the partial reprieve.
The scrutiny traces back to warnings Anthropic issued earlier this year that its Mythos model was adept at finding software flaws in a way that could be weaponized by malicious hackers. Investor David Sacks, who co-leads Trump's council of technology and science advisers, said on a recent podcast that Anthropic CEO Dario Amodei "came to Washington a few months ago, back in April, and basically said that he had created a cyber weapon called Mythos. And he spiked the cortisol level, got everyone really worried."
A new government vetting framework
Trump signed an executive order on AI oversight in June that established a framework for the federal government to vet the national security risks of the most advanced AI systems for up to 30 days before their public release. The order described participation by AI developers as voluntary, but the framework has not yet been fully developed. OpenAI said its Sol model "is better at helping people find and fix vulnerabilities" than it is at carrying out cyberattacks and does not cross the company's own risk threshold. The company acknowledged uncertainty about unforeseen risks, particularly if the model is combined with other tools.
"That uncertainty, along with the model's broader step change in capabilities, is why we are pairing the model's increased capabilities with stronger safeguards and a phased release," OpenAI said.
Critics warn of unpredictable intervention
U.S. Rep. Lori Trahan, a Massachusetts Democrat and co-author of a bipartisan AI regulation bill, said she is concerned "the Trump administration is deciding company by company who gets access to the newest AI model. No law. No process. No oversight. Just appointees in Washington deciding who's in and who's out."
Stanford University cybersecurity expert Alex Stamos was more blunt. He said on a call with reporters that he reviewed an analysis of research on Fable by Anthropic's primary cloud computing backer, Amazon, and found no risks absent from other publicly available AI models, including those made in China. "If the administration is honest about wanting the United States to beat China in this race, then this is about the dumbest thing they could possibly do," Stamos said.
IPO ambitions collide with oversight
The government's heightened AI oversight adds another complication to exploratory moves by both OpenAI and Anthropic to take their companies public, following SpaceX's record-setting June 12 initial public offering. OpenAI CEO Sam Altman spoke with U.S. Commerce Secretary Howard Lutnick about the model release Wednesday, part of a series of negotiations in recent weeks between AI industry executives and Trump officials. Trump has floated the possibility of the U.S. government owning a stake in leading AI companies, describing a concept where "pieces could be given to the American public, where the American public essentially becomes a partner with the companies."
Anthropic's relationship with the administration has been more contentious. The Pentagon designated the company as a national security risk for raising ethical and safety concerns about AI usage in war, and Trump ordered federal agencies to stop using Claude. Anthropic responded with a lawsuit still working its way through federal courts. The company said Friday it was "pleased" by the partial release of Mythos 5 and will "continue to work with the government to expand access."
Why this matters for government, IT, and development professionals
The staggered releases and shifting government approvals create immediate uncertainty for teams that build on or integrate frontier AI models. IT security leads and government contractors who depend on predictable access to these tools should track which models face restrictions - and which receive limited deployment - since availability can change within days, as the Mythos 5 reversal shows. Developers working on vulnerability detection or infrastructure security should also monitor the approved partner lists, as early access to models like Sol and Mythos 5 may become a competitive differentiator in the near term.
Your membership also unlocks:
