Anthropic’s Mythos 5 Surges Back into Restricted Access After US Govt Clears AI Model Amid Security Feud

Anthropic’s Mythos 5 Surges Back into Restricted Access After US Govt Clears AI Model Amid Security Feud​

Breakthrough in AI Regulation as Government Approves Limited Release of Mythos 5​

Anthropic PBC has been granted partial restoration of access for its advanced artificial intelligence model, Mythos 5. The clearance follows intensive review by the U.S. government to resolve national security concerns regarding the technology’s potential threats. This approval allows Anthropic to begin deploying the powerful system to a selected group of trusted partners and cyber defenders.

Commerce Secretary Howard Lutnick confirmed that Anthropic has made "significant progress" working with the government to mitigate risks associated with the Covered Models. The resolution eases a period of intense confrontation initiated two weeks prior when federal regulators abruptly restricted access to Mythos 5 and Fable 5 for foreign nationals due to concerns over guardrail circumvention.

Security Clearance and Redeployment Status​

Anthropic announced that the US government has cleared Mythos 5, confirming its redeployment to a small cohort of cyber defenders and infrastructure providers. The company is working diligently to provision these approved partners and restore their access as rapidly as possible.

The clearance marks a crucial pivot point in the ongoing regulatory battle surrounding advanced generative AI. While the availability of Mythos 5 has been restored for specific uses, Commerce Secretary Lutnick noted that the government remains watchful over the security elements of the technology.

Ongoing Tensions and Rival Industry Impact​

The federal government is still restricting certain aspects of Fable 5, a related model developed by Anthropic. The company plans to continue working with the government to bring general access for Fable 5 back soon, although specific details remain private following ongoing discussions.

This high-stakes regulatory intervention into AI operations reverberated throughout the technology sector. Rival startup OpenAI subsequently limited its rollout of GPT-5.6 in response to pressure from the Trump administration, restricting a preview version to select U.S. government approved partners.

The Wider Feud and Corporate Resistance​

The current showdown over Anthropic's models is part of a deeper, months long conflict with the Pentagon. Previously, after contract discussions failed, the Defense Department declared Anthropic a supply-chain risk, initiating efforts to find alternative AI providers for military use.

Anthropic privately held the company positioned itself as a responsible developer, but it faced serious challenges regarding its dual capabilities. The firm expressed strong objections via a blog post when its models were initially disabled by the government. The company cautioned that imposing such standards across the industry would "essentially halt all new model deployments for all frontier model providers."

Industry Stakes and Future Policy Frameworks​

The incident brought into sharp focus emerging security concerns regarding cutting edge AI capabilities, presenting a significant challenge to Anthropic weeks after it filed confidentially for an initial public offering with a valuation exceeding $900 billion.

Senior executives, including co-founder Tom Brown, met with Lutnick and other administration officials during the negotiations. A person familiar with the private talks stated that US officials mandated that restrictions would be lifted once all security concerns had been fully addressed by Anthropic. The company is also committed to working with the government to establish a future policy framework for managing such technological conflicts.