# OpenAI and Anthropic Navigate Unprecedented Government Oversight of Advanced AI Models
The landscape of artificial intelligence development in the United States has entered uncharted territory, as both OpenAI and Anthropic find their most powerful new models subject to direct government vetting and release restrictions at the request of the Trump administration.
## OpenAI Restricts GPT-5.6 Sol Release
OpenAI announced on Friday that it is limiting the rollout of its latest AI model, GPT-5.6 Sol, following a request from President Donald Trump’s administration. The company stated that the new product would be accessible only to customers specifically approved by the government, marking what appears to be an extraordinary level of federal involvement in the release of commercial AI technology.
In a public statement, OpenAI expressed reservations about the arrangement becoming a permanent fixture. “We don’t believe this kind of government access process should become the long-term default,” the company said, characterizing the testing period as a temporary measure on the “path to broader availability in the coming weeks.”
OpenAI described the Sol model (pronounced “SOHL,” like the Spanish word for sun) as being “better at helping people find and fix vulnerabilities” than at carrying out cyberattacks, and said it does not cross the company’s own internal risk threshold. However, the company acknowledged that unforeseen risks could emerge, 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 Friday. The company has not publicly identified any of the approximately 20 customers approved to use the new model so far.
## Anthropic’s Partial Victory on Mythos 5
Hours after OpenAI’s announcement, rival AI firm Anthropic revealed that the Trump administration had approved a limited release of its strongest cybersecurity model, Mythos 5. The approval came roughly two weeks after the U.S. Commerce Department effectively banned the model’s use.
Anthropic said Mythos 5 would be “redeployed to a small group of cyber defenders and infrastructure providers.” The company stated it was “pleased” by the partial release and would “continue to work with the government to expand access” and make its other restricted model, Fable 5, available again to general users.
U.S. Commerce Secretary Howard Lutnick communicated to Anthropic in a letter dated Friday that the company’s efforts to address the government’s concerns “yielded significant progress.”
## The Road to Government Scrutiny
The heightened oversight traces back to warnings Anthropic issued earlier this year about its Mythos model’s ability to identify software vulnerabilities in ways that could potentially be weaponized by malicious actors, posing a threat to critical computer networks worldwide.
In response, Trump signed an executive order in June establishing a framework for the federal government to evaluate the national security risks of the most advanced AI systems for up to 30 days before their public release. While the order described participation by AI developers as voluntary, the framework has not yet been fully developed.
Some of Trump’s allies have pointed fingers at Anthropic and its CEO, Dario Amodei, for triggering the increased scrutiny. Investor David Sacks, who co-leads Trump’s council of technology and science advisers, said on a recent podcast: “Dario 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. And there was some truth to it in terms of the sense that this model had advanced cyber capabilities.”
## Critics Sound the Alarm
The government’s intervention has drawn sharp criticism from lawmakers and technology experts alike. U.S. Rep. Lori Trahan, a Massachusetts Democrat and co-author of a bipartisan AI regulation bill, voiced concern over the ad hoc nature of the process. “The Trump administration is deciding company by company who gets access to the newest AI model,” she said. “No law. No process. No oversight. Just appointees in Washington deciding who’s in and who’s out.”
A broad coalition of technology experts has also challenged the government’s actions, particularly regarding Anthropic’s Fable 5 model, which was taken offline and remained unavailable even after restrictions on the more powerful Mythos 5 were lifted on Friday.
Stanford University cybersecurity expert Alex Stamos was blunt in his assessment. “I just want to say that pretty much nobody in the cybersecurity industry believes that there’s any factual basis for this action,” he said during a call with reporters earlier in the week.
Stamos, who serves as chief product officer at AI security company Corridor and is a former chief security officer at Meta, reviewed an analysis of Fable 5 conducted by Anthropic’s primary cloud computing partner, Amazon. He said he found no risks that aren’t already present in other publicly available AI models, including those developed 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.
## Oversight Intensifies as Companies Eye Public Markets
The government’s increased AI oversight adds a layer of complexity to the plans of both OpenAI and Anthropic as they explore initial public offerings on Wall Street, following SpaceX’s record-setting June 12 IPO.
OpenAI CEO Sam Altman spoke with Commerce Secretary Howard Lutnick about the model release on Wednesday, part of ongoing negotiations between AI industry executives and Trump administration officials in recent weeks.
Anthropic has also participated in those discussions, though Amodei’s relationship with the Trump administration has been notably more contentious. The Pentagon previously designated Anthropic as a national security risk after the company raised ethical and safety concerns about AI usage in warfare. Trump himself ordered federal agencies to stop using Anthropic’s Claude chatbot, prompting the company to file a lawsuit that remains pending in federal courts.
Meanwhile, Trump has floated the possibility of the U.S. government taking ownership stakes 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.”
The White House said Friday that it continues to collaborate with frontier AI labs on addressing the challenges of scaling the rapidly advancing technology.
—
*This article is based on reporting from the original source.*
