Mythos 5 Reopened to Over 100 U.S. Institutions, Fable 5 Restrictions Still Not Lifted
N.R. Finch
The U.S. Commerce Department confirmed that Anthropic's Mythos 5 has been restored to over 100 organizations, but consumer-facing Fable 5 remains restricted — meaning Anthropic's commercial recovery is only half-done, while government sign-off on frontier AI releases is fast becoming the new normal.
What exactly was unlocked — and what wasn't?
Commerce Secretary Howard Lutnick wrote to Anthropic co-founder Tom Brown confirming that Mythos 5 can reopen to more than 100 U.S. organizations, spanning large enterprises and government agencies.
Approved organizations may also grant access to their foreign-national employees; Anthropic's own foreign-national staff are included.
But Fable 5 — the consumer-facing version Anthropic built on top of Mythos with additional safety guardrails — remains restricted, with no timeline for restoration.
This means → Anthropic has regained the right to serve major clients, but the door to ordinary users is still shut.
Why was access restricted in the first place?
According to WIRED, the Trump administration's concerns stemmed from Anthropic sharing Mythos access with a South Korean telecom firm believed to have ties to China.
Subsequently, Amazon and the NSA discovered vulnerabilities in Fable 5 that could bypass its safety guardrails, prompting further government action.
In plain terms = one problem was giving the model to the wrong partner; the other was a hole in the model's own safety wall. Both together triggered the White House's June 12 restriction order.
What did Anthropic do to regain access?
Anthropic sent senior cybersecurity and AI-safety staff to Washington to meet government officials; public-policy lead Sarah Heck spearheaded negotiations with the Commerce Department.
Lutnick's letter stated that Anthropic's collaboration with the government had achieved "significant progress" and that "appropriate safeguards are now in place."
Yet the letter also made clear: all other requirements from the original June 12 directive remain in effect. This reflects a partial opening, not a full all-clear.
What does this mean for the broader AI industry?
Dean Ball, a former White House AI adviser now at OpenAI's strategic-futures team, wrote on June 16 that this clash has shown frontier AI developers they "now need an explicit green light from the government."
That same week, OpenAI announced it would delay the release of its upcoming GPT-5.6 series at the government's request — both companies' experiences point to the same trend.
This means → U.S. government intervention in the release cadence of frontier AI models is deepening systematically. The era of "build it first, ship it later" may already be over.
How big is the commercial cost to Anthropic?
Anthropic had already sued the Trump administration over supply-chain risk designations, after the government tried to set red lines on military contractors' use of its AI models.
Whether Fable 5 ultimately gets cleared for public access is the key test of whether Anthropic's commercialization can fully recover.
In plain terms = the lawsuit is still ongoing, the consumer product is still locked — for a company in a high-growth phase, every additional day of restriction means lost revenue and rising customer-attrition risk.
Content is for reference only, not financial advice.