U.S. Government Orders Anthropic to Disable Claude Fable 5 and Mythos 5 - Three Days After Launch

For the first time, the U.S. government has forced an American AI company to pull its flagship models off the market. On Friday, June 12, 2026, Anthropic said it received a federal export-control directive - citing national-security authority - that required it to abruptly disable its two most capable models, Claude Fable 5 and Claude Mythos 5. Fable 5 had been publicly available for just three days. By the weekend, both were switched off for every user worldwide.

This is a genuinely landmark moment in AI policy, and the arguments on both sides are serious. Here is exactly what happened, what the government is concerned about, how Anthropic is responding, and why every frontier AI lab is now watching closely.

1. What the Directive Says

According to Anthropic’s public statement, the U.S. government, “citing national security authorities, has issued an export control directive” covering Fable 5 and Mythos 5. The order applies to “any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.”

That scope is the key to why the impact is global. Anthropic says it has no way to reliably separate foreign nationals from U.S. citizens across its live products in real time, so the only way to comply was to shut the models off entirely: “The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.” Crucially, the company added that “access to all other Anthropic models will not be affected” - so Claude Opus 4.8 and the rest of the lineup stayed online.

2. What Fable 5 and Mythos 5 Are

The two models sit at the very top of Anthropic’s range, and understanding the difference explains why the government focused on them:

• Mythos 5 is the more powerful of the pair - a frontier model that, according to reporting, was previewed in April 2026 and made available only to a small number of vetted organizations because of its exceptional ability to find software security vulnerabilities.
• Fable 5, released June 9, is the guardrailed version built for public release - the same underlying capability with hard blocks on high-risk areas such as cybersecurity and biology. At launch it was described as the most capable AI model available to the general public.

In other words, the very capability that made these models impressive - finding and reasoning about flaws in code - is also what put them in the government’s crosshairs.

3. The Government’s Concern

Anthropic says its understanding is that the underlying issue is a claimed method of bypassing - or “jailbreaking” - Fable 5’s safeguards. In practical terms, the concern centers on prompting the model in ways that get it to identify software flaws in specific codebases, despite the guardrails meant to limit that.

From a national-security standpoint, the worry is straightforward: a widely available, highly capable tool for discovering software vulnerabilities could be misused by hostile actors. Export-control authority is the legal mechanism the government has reached for to limit who can access that capability.

4. Anthropic’s Response

Anthropic’s public pushback is unusually pointed for a company responding to its own regulator. It makes three arguments:

• The vulnerability is minor. The company says it reviewed “a small number of previously known, minor vulnerabilities” that “appear relatively simple,” and characterizes the problem as “a narrow, non-universal jailbreak” rather than a fundamental failure of its safeguards.
• The capability isn’t unique. Anthropic argues the ability in question “is widely available from other models (including OpenAI’s GPT-5.5), and is used every day by the defenders who keep systems safe” - i.e., it’s a standard tool for legitimate cybersecurity work, not a novel weapon.
• The remedy is disproportionate. Its bottom 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.”

5. Why It Matters

A new precedent. This appears to be the first time the U.S. has used export-control authority to force a commercial AI model offline. However the specific dispute is resolved, the tool now exists in the regulatory playbook - and that changes the risk calculus for every company shipping frontier models.

The whole industry is exposed. If a narrow, contested vulnerability can pull a model used by hundreds of millions of people on a few days’ notice, every frontier lab has to weigh launch speed against compliance and national-security risk in a way it didn’t before.

An irony worth noting. Anthropic has positioned itself as the safety-first lab, repeatedly warning that frontier AI is becoming dangerously capable. That candor is part of its brand - but it may also have helped invite the very scrutiny now disrupting its business, just weeks after the company confidentially filed to go public. A messaging strategy built on emphasizing capability and risk can cut both ways.

What We Still Don’t Know

• How long the suspension lasts. Anthropic has framed the disabling as necessary for compliance, but the path back - appeal, a technical fix, a narrowed order - isn’t yet public.
• The exact legal basis. The directive is described as an export-control action citing national-security authority; the precise statutory mechanism and process have not been fully detailed publicly.
• Whether other labs are next. If the concern is a class of capability rather than one model, comparable models elsewhere could face similar scrutiny.

The Bottom Line

A frontier AI model that was the most capable public system available on Tuesday was switched off by Friday - not because it failed, but because of a dispute over a narrow capability and the authority to govern it. Anthropic’s other models keep running, and the company is pushing back hard. But the bigger story is the precedent: the U.S. government has now shown it will reach for export-control authority to pull an AI model from the market, and the entire industry will be calibrating to that new reality.

Sources

• Anthropic - Statement on the US government directive to suspend access to Fable 5 and Mythos 5 (primary)
• CNBC - Anthropic disables access to Fable 5 and Mythos 5 to comply with government directive
• TechCrunch - Anthropic’s safety warnings may have just backfired
• NBC News - Anthropic suspends new AI models after government directive
• 9to5Mac - Anthropic pulls Claude Mythos 5 and Fable 5 following US government directive

Curated by Jerry Cards - jerrycards.com. We research the week’s most consequential tech, science, and business news so you don’t have to. More at jerrycards.com/news.