US Clears Anthropic's Mythos 5 for Cyber Defenders

On 27 June 2026, the US government cleared Anthropic to redeploy Mythos 5, which the company calls its strongest cybersecurity model, to organizations that operate and defend US critical infrastructure. The move partly reverses a block imposed on 12 June. Fable 5, Anthropic's public-facing model, stays offline, and access to Mythos 5 now runs through a vetted set of partners chosen with the government.

What the government cleared

On 27 June, Anthropic said the government had notified it that Mythos 5 could be redeployed to a set of US organizations that operate and defend critical infrastructure, and that it was restoring access quickly.

Items labeled projection are forecasts or stated plans, not built facts.

Press accounts (Semafor and Reuters, via TechCrunch) put the cleared group at more than 100 US agencies and companies, and report that non-American staff at approved organizations, including some of Anthropic's own employees, may now use the model.

Anthropic's own wording was narrower than the press figure. It described "a set" of organizations rather than a count, so the gap between that phrase and "more than 100" is worth holding lightly until a primary list appears.

"Mythos 5, our strongest cybersecurity model, can be redeployed to a set of US organizations that operate and defend critical infrastructure."

Anthropic, 27 June 2026

The block it partly reverses

The clearance unwinds part of a 12 June export-control order that cut off both Mythos 5 and Fable 5. That order, which Anthropic said it received at 5:21pm ET, required suspending access for all foreign nationals, including Anthropic's own foreign-national employees, after the government raised concern about a reported method of jailbreaking Fable 5 to find software vulnerabilities. Anthropic disputed the basis, calling it a narrow, non-universal jailbreak.

Before the block, Mythos 5 ran through Anthropic's Project Glasswing, an invite-only cyber-defense program. Estimates of its size vary: Bloomberg put it at about 200 firms, including Apple, Google, Nvidia, Cisco, and Microsoft, while Reuters described roughly 100 companies and institutions. The 27 June decision restores a portion of that reach, on tighter terms, while leaving the public model dark.

The letter, and the law behind it

On the government side, reporting describes a letter from Commerce Secretary Howard Lutnick to Anthropic, dated Friday 26 June and seen by Bloomberg, citing "significant progress" in the company's work with the government to address risks associated with what it called "Covered Models." Per Reuters, the letter lifts the export-license requirement for Mythos 5 at approved companies, including their non-US employees and Anthropic's own, while keeping restrictions in place for firms not on the approved list. A Commerce spokesman said the department had worked to keep the US the global leader in AI while safeguarding national security.

The letter reportedly refers to "Covered Models," language that maps to Executive Order 14409 of 2 June 2026 and its "covered frontier model" concept. That order directs agencies to build a classified process for designating such models, with a 60-day deadline that runs to roughly 1 August Projection.

Section 3(c) of that order is explicit that nothing in it authorizes a mandatory licensing or preclearance requirement for releasing AI models. The Mythos 5 decision, made one model and one set of partners at a time, functions as a case-by-case approval in the gap before the formal framework exists.

What stays unresolved

Fable 5 is the open question. The Lutnick letter reportedly does not mention it, and Anthropic said it is still working with the government to make Fable 5 available for general use again Projection. People familiar with the talks said discussions were continuing toward that goal Projection.

For now, the most capable US models reach users case by case, decided in Washington. Anthropic, in the post announcing the shutdown, warned that a blanket version of the standard would "essentially halt all new model deployments for all frontier model providers." The Mythos 5 clearance shows the narrower version of that standard already at work.

Sources

1. Anthropic, statement on Mythos 5 redeployment (via Fortune, CNN, and TechCrunch), 27 June 2026.
2. Anthropic, "Statement on the US government directive to suspend access to Fable 5 and Mythos 5", 12 June 2026.
3. The White House, "Promoting Advanced Artificial Intelligence Innovation and Security" (Executive Order 14409), 2 June 2026.
4. Fortune, "Anthropic's Mythos 5 AI model cleared by U.S. for wider use", 27 June 2026.
5. TechCrunch, "Trump Admin releases Anthropic Mythos to be used by more than 100 US companies, agencies", 26 June 2026.
6. The Next Web, "US clears Anthropic to restore Mythos 5 access for trusted cyber defenders", 27 June 2026.
7. CNN Business, "US government allows Anthropic limited release of AI model that sparked cybersecurity concerns", 27 June 2026.