Full Breakdown
U.S. Government Restores Limited Access to Anthropic’s Mythos 5 AI Model
6/27/2026, 9:47:40 PM
Limited Release of Mythos 5
On 26 June 2026 the Commerce Department notified Anthropic that Claude Mythos 5 could be redeployed to a “trusted-partner” list of roughly 100 U.S. companies and agencies, reversing the 12 June export-control order that halted Mythos 5 and its consumer version, Claude Fable 5.
Background & Context
Mythos 5, launched 9 June, is a cybersecurity-focused LLM with a 1-million-token context window. Within days the Commerce Department cited jailbreak risks, foreign-intelligence misuse, and a link to a telecom, prompting export controls.
Timeline
- 9 Jun 2026 – Anthropic launches Mythos 5 and Fable 5.
- 12 Jun 2026 – Export-control order halts access.
- 26 Jun 2026 – Lutnick authorizes limited Mythos 5 access.
- 27 Jun 2026 – Reports of imminent Fable 5 lift.
Key Figures & Groups
Anthropic – chief compute officer Tom Brown, CEO Dario Amodei. Howard Lutnick – U.S. Commerce Secretary. Benno Kass – Commerce spokesperson. John Coleman – legislative counsel, Foundation for Individual Rights and Expression. Sam Altman – OpenAI CEO.
Data & Statistics
Access granted to over 100 entities, including Fortune 500 firms. Mythos 5 is Anthropic’s strongest cybersecurity model. No list of approved partners exists.
Official Statements & Responses
Lutnick wrote that Anthropic had addressed risks to the Covered Models and that safeguards now permit trusted partners to use Mythos 5. Anthropic’s statement highlighted rapid restoration for approved providers and a commitment to make Fable 5 generally available. A Commerce spokesperson noted the two-week effort to keep the United States “the global leader in AI while safeguarding our security.”
Criticism & Opposition
John Coleman warned that “no one knows how these companies are picked and why everyone else is excluded,” and Sam Altman said he “doesn’t like the idea of the government picking the customers.” Critics argue the process concentrates power and lacks transparency.
Conflicting Reports & Gaps
Sources differ on when Fable 5 will be restored—some cite an imminent lift, others call the schedule unclear. The criteria for “trusted partners” and the exact list of approved entities remain undisclosed.
Why It Matters / Impact
Dual-use risk of AI is evident: tools for cybersecurity and drug design can enable attacks. Treating models as controlled sets a precedent that may shape U.S. AI policy, affect competition with China, and dictate how firms negotiate safeguards.
What’s Next
Anthropic expects talks to restore Fable 5, while the Commerce Department drafts vetting rules for frontier AI. Industry awaits guidance on export-control criteria and legislation to formalize the process.
Verbatim Quotes
- “I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model.” — Howard Lutnick, U.S. Commerce Secretary
- “Anthropic has worked with the U.S. government to address risks associated with the Covered Models. These efforts have yielded significant progress,” — Howard Lutnick, U.S. Commerce Secretary
- “We are pleased to see this progress and continue to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again.” — Eduardo Maia Silva, Anthropic spokesperson
- “No one knows how these companies are picked and why everyone else is excluded,” — John Coleman, Legislative Counsel, Foundation for Individual Rights and Expression
