At 5:21 p.m. ET on Friday, the U.S. Commerce Department sent Anthropic CEO Dario Amodei a letter barring distribution of Claude Fable 5 and Claude Mythos 5 to any foreign national, citing national security authorities. Within hours Anthropic disabled both models for every customer worldwide, because the company can’t operationally separate foreign nationals from anyone else, including its own non-citizen staff. Fable 5 had been live for three days. Per Vals AI’s benchmarks, it was the most capable publicly available model on the market when Commerce Secretary Howard Lutnick’s directive arrived.
The rest of the Claude lineup, including Opus 4.8, is unaffected. The two killed models are precisely the ones Anthropic had spent months telling Washington were dangerous.
That’s the structural irony driving the entire story. Mythos 5 had been previewed in early April under Project Glasswing, a controlled-sharing program offering the model to roughly 50 vetted organizations, Amazon, Apple, Google, Microsoft, and CrowdStrike among them, for defensive cybersecurity work. By Anthropic’s own account, Mythos 5 had identified vulnerabilities in every major operating system and web browser it was tested against. The marketing was the warning, and the warning became the pretext.
Anthropic’s statement makes the grievance explicit. The company says it reviewed a demonstration of the technique cited verbally by officials (the letter itself didn’t specify) and characterizes it as “a potential narrow, non-universal jailbreak,” involving “a small number of previously known, minor vulnerabilities” that are “widely available from other models (including OpenAI’s GPT-5.5), and is used every day by the defenders who keep systems safe.” Then the harder line: “We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people.” And the implication that escalates this from a compliance story to an industry one: “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”
The political backdrop matters. In early March the Pentagon designated Anthropic a “supply chain risk,” barring defense contractors from using Claude on government work; Anthropic is challenging the label in federal court. Trump technology advisors including former AI and crypto czar David Sacks and Pentagon research undersecretary Emil Michael have publicly attacked the company. In April, OpenAI’s Sam Altman told podcaster Ashlee Vance that Anthropic’s handling of Mythos amounted to “fear-based marketing.” Earlier this month Anthropic filed a confidential IPO at a $965 billion valuation on a $47 billion revenue run rate.
The lab built its brand on telling regulators which capabilities deserve alarm. The regulators listened, then chose which company to apply the alarm to.
Sources
- https://www.anthropic.com/news/fable-mythos-access
- https://www.bloomberg.com/news/articles/2026-06-13/anthropic-says-us-limits-foreign-access-to-fable-5-mythos-5
- https://www.cnbc.com/2026/06/12/anthropic-disables-access-to-fable-5-and-mythos-5-to-comply-with-government-directive.html
- https://techcrunch.com/2026/06/12/anthropics-safety-warnings-may-have-just-backfired-the-government-has-pulled-the-plug-on-its-most-powerful-ai/
- https://fortune.com/2026/06/13/anthropic-disables-fable-mythos-export-controls-national-security-threat/