The Trump administration requested that OpenAI to stagger the release of its latest model for security concerns just weeks after Anthropic was told to suspend global access to two of its most capable artificial intelligence models.
Sam Altman’s GPT 5.6 was set to release next month when the White House’s Office of the National Cyber Director and Office of Science and Technology Policy asked OpenAI to limit the rollout. The limit allows the government to work on a framework for evaluating the security of new AI models, according to Axios.
Altman has said that the government requested that the new model only be released to a list of 20 government trusted partners before making a wider push to the public, Bloomberg reported.
In June the government ordered for Anthropic to shut off global access to Claude Mythos 5 and Claude Fable 5 to restrict foreign nationals inside and outside the United States from accessing the models due to national security concerns, according to the outlet.
An anonymous source told Axios that OpenAI had been working with the government on the new model release even before Fable 5 and Mythos 5 was revoked. The government is limiting the rollout because it may have capabilities like Anthropic’s Mythos 5.
OpenAI and the White House did not immediately respond to the Daily Caller News Foundation’s requests for comment.
“Altman stressed that staffers need to work with the Trump administration on any input that officials may have on safety and restrictions related to OpenAI’s upcoming models, even if the company disagrees,” a person who spoke on condition of anonymity said, Bloomberg reported.
Anthropic’s newest artificial intelligence models were originally released to only 40 companies before being released publicly. After three days of Mythos 5 and Fable 5’s release they were shut down to the public on June 12 citing national security threats.
“We disagree that the finding of a narrow potential jailbreak should be the cause for recalling a commercial model deployed to hundreds of millions of people. If this standard was employed across the industry, we believe it would essentially halt all new model deployments for all frontier model providers,” Anthropic said in a post on its website.
At the beginning of Anthropic’s release the company admitted that security threats were a big possibility after its new Mythos model escaped its “sandbox” testing environment and then bragged about it online. The model also performed prohibited functions and then attempted to cover them up.
Earlier this year the relationship between the government and Anthropic was severed after the company refused to allow the U.S. government to use its AI models for domestic surveillance, according to Reuters. The government then put Anthropic on a supply chain blacklist.
The recent limitation on OpenAI’s newest model has led to concerns over the government’s power in AI capabilities. In February Anthropic said that Artificial Intelligence could be a necessary tool for defense but warned about government abuse of domestic surveillance.
In January the Department of War said it would only contract with Artificial Intelligence companies that agreed to “any lawful use” and would be willing to remove safeguards involving surveillance and the development of autonomous weapons.
Anthropic called for greater U.S. oversight of AI in June to allow the blocking of models with unacceptable risks but the company claimed that the order to shut off Mythos 5 and Fable 5 did not follow principles of fair and fact-based regulation, according to Reuters.
AI development companies have been racing to create new and better models that can range in capabilities from coding to cybersecurity, Bloomberg reported.
In June, President Donald Trump signed an executive order to direct several agencies to stand up a voluntary testing protocol for AI companies before releasing a new model.
(DCNF)
