US Government Orders Anthropic to Pull Latest AI Models Within 90 Minutes

Anthropic’s Claude Fable 5 and Mythos 5 are its flagship generative AI models. Global access had been suspended for 18 days after the US government raised national security concerns. The episode affected the cross-border availability of advanced AI technology and highlighted the direct impact of model-safety and export-control policies on businesses and users.

The US Commerce Department lifted export controls on Claude Fable 5 and Mythos 5 on June 30. Anthropic subsequently confirmed that it would begin gradually restoring access for users worldwide on July 1. The company also strengthened its safety checks and began developing an AI jailbreak evaluation framework to address concerns about model misuse.

Anthropic has positioned Claude Mythos 5 as its most powerful cybersecurity model, making its capabilities and access controls consequential for critical infrastructure protection and AI governance. With the public-facing Fable 5 still blocked, the gap between government and civilian access to advanced models has fueled controversy over technological monopolies and double standards.

Anthropic recently announced that Mythos 5 had been cleared to come back online, but only for certain US critical infrastructure organizations, with no access for the general public. Fable 5 remains blocked. Anthropic did not disclose the exact announcement date, the number of authorized organizations or any financial amounts involved, prompting criticism from the online and open-source communities that it is using AI risks to entrench national and technological advantages.

Anthropic’s Mythos 5 can automatically identify software vulnerabilities. The White House had previously opposed giving another 70 companies access, citing national security risks and the allocation of computing resources. The case highlights the Trump administration’s balancing act between the AI arms race, control of critical computing capacity and cooperation with frontier AI companies. Reports did not disclose any transaction value.

The latest reports indicate that the U.S. government has reversed course and allowed about 100 trusted companies to use Anthropic’s Mythos 5, broadening access beyond the additional 70 companies previously blocked. However, available information did not specify the exact approval date, participating companies, usage fees or computing quotas. Details of the national security review requirements have yet to be announced.

Anthropic's Mythos and Fable are AI models designed for advanced reasoning and cybersecurity tasks. The tests involved the defenses of classified U.S. National Security Agency networks. A model capable of quickly identifying weaknesses in national-security systems could affect AI safety governance, government procurement and the Five Eyes alliance's plans for technological sovereignty. As of July 19, 2026, no publicly disclosed amount was associated with the matter.

Recent reports said Mythos penetrated NSA systems within hours during a government red-team test, prompting U.S. lawmakers to call for mandatory third-party testing. Foreign media later clarified that the test took place in a simulated environment and did not involve an actual intrusion into the NSA's classified networks. The U.S. government has also restricted accounts belonging to non-U.S. nationals from accessing Mythos and Fable. As of July 19, 2026, officials had not disclosed the test date, full results or the date the restrictions took effect.

The U.S. Commerce Department has previously sought to constrain advanced AI development largely through controls on chips and equipment. This marks the first time it has extended export restrictions directly to software models. Fable 5 and Mythos 5 were targeted because their advanced cybersecurity reasoning capabilities were seen as potentially useful to foreign nationals seeking to exploit vulnerabilities, establishing a policy precedent.

The Commerce Department sent Anthropic a letter on June 12 barring foreign nationals from accessing Fable 5 and Mythos 5, forcing the company to suspend the services worldwide. Telecommunications operators in Europe, Canada and South Korea were also affected. Anthropic has sent senior executives to the White House for talks, but Mythos 5 remains offline. Officials have not announced a timetable for lifting the restrictions or disclosed any amount involved.

Anthropic’s Mythos AI model is positioned as a system with advanced cybersecurity capabilities. But its powerful offensive and defensive tools could also be misused, drawing scrutiny from the European Commission and financial regulators. Anthropic has pledged to comply with the EU’s AI Code of Practice, assess the model’s risks and take steps to mitigate them.

EU officials have now met with Anthropic for a briefing on cybersecurity concerns surrounding Mythos, while euro-area finance ministers have separately raised requirements concerning bank access. Available information does not disclose the exact date of the meeting, the number of banks affected or any transaction amounts. Attention will now turn to access permissions and risk-control standards.

Anthropic’s new Claude Mythos model has demonstrated a strong ability to identify software vulnerabilities. It could help companies and researchers patch systems, but could also be misused to rapidly find weaknesses and expand cyberattacks. The issue is shaping the direction of U.S. AI regulation and testing the balance between model innovation and national cybersecurity.

The White House is concerned that Claude Mythos could be used in large-scale cyberattacks, according to the latest reports. The Trump administration is considering moving away from its relatively hands-off stance and requiring new AI models to undergo government security reviews and standardized testing before release. No implementation date, thresholds, responsible agency or related budget has been announced.

Anthropic’s relationship with the US government had become strained after it rejected the Department of Defense’s demand for an “unrestricted license” to its AI technology and was placed on a blacklist. CEO Dario Amodei’s return to the White House could determine whether its latest model, Mythos, can be deployed for government and national security applications. The talks also bear on the policy boundaries AI suppliers set for military uses.

Amodei met the White House chief of staff and Treasury Secretary Scott Bessent to discuss Mythos deployment and national security, marking the first senior-level effort to ease tensions since the dispute. The Pentagon’s chief technology officer later said Anthropic remained on the blacklist but that Mythos would receive an exception. Reports did not disclose the exact meeting date, procurement value or deployment timetable.

Anthropic was designated a supply-chain risk on February 27 after refusing to allow the U.S. Defense Department to use Claude for “all lawful purposes,” insisting on exclusions for mass domestic surveillance and fully autonomous weapons. Trump also ordered federal agencies to stop using the model, giving existing users a six-month grace period. The dispute affects a $200 million defense contract and highlights the tension between national-security demand for frontier AI and safety limits.

Anthropic launched Project Glass Wings on April 7, giving more than 40 organizations access to Mythos Preview and pledging $100 million in usage credits and $4 million in donations. The model has already identified thousands of high-severity vulnerabilities. Bloomberg reported on April 16 that the White House was considering allowing major federal agencies to use it with safeguards. The CEO met the White House chief of staff and Treasury secretary on April 17, while Axios disclosed on April 19 that the National Security Agency had already been using the model.

The Trump administration blacklisted Anthropic and restricted federal agencies from using its technology after the company insisted that AI should not be deployed in weapons systems. Yet Anthropic’s models remain strategically valuable in cybersecurity and software development and are still being evaluated internally, highlighting a policy conflict between national security needs and AI safety principles.

US media recently reported that the Commerce Department, congressional committees and some federal officials had circumvented the ban to privately test the cyberattack and coding capabilities of Anthropic’s most powerful model, Mythos, and receive related technical briefings. The reports did not disclose the exact dates of the tests, the committees involved or any associated amounts. The restrictions arising from Anthropic’s blacklisting remain in place.