Anthropic adds new security measures to get restrictions lifted on Fable 5 and Mythos 5

The Trump administration lifted restrictions on two of Anthropic's flagship models, Fable 5 and Mythos 5, in early July 2026. The condition: the company had to introduce new security measures meant to satisfy the federal government's requirements. Wired reported the news.

How Anthropic ended up under restrictions

Tension between Anthropic and US federal authorities had been building as the state tightened control over advanced AI systems. The Trump administration has consistently pushed for more influence over how the most powerful models are distributed, both at home and abroad. AI systems are increasingly treated as strategic assets that require national oversight on a par with military technology.

Fable 5 and Mythos 5 are Anthropic's current-generation flagship models. The restrictions on their use effectively cut the company off from government contracts and from large corporate clients that operate under federal security standards. For a company that positions itself as a "safety-first" AI lab, losing access to the US government market was a serious operational blow.

What is the price of unblocking?

The details of the agreement are confidential, but Wired describes the situation unambiguously: the restrictions were lifted "not without conditions" (strings attached). Anthropic has added new control mechanisms, reportedly covering usage monitoring, customer verification and limits on access from certain jurisdictions.

Key facts from Wired's report:

• Restrictions were lifted on Fable 5 and Mythos 5, two of Anthropic's most advanced systems
• The company introduced new security measures as a condition of the agreement with the government
• The specific requirements set by the administration have not been publicly disclosed
• Anthropic has mended relations with the White House after a period of tension

The "access in exchange for oversight" model is becoming standard practice in the US AI industry. OpenAI and Google DeepMind went down a similar path: they signed agreements with federal agencies and launched special "government" versions of their products, receiving contracts with the Pentagon and intelligence agencies in return.

Why this matters right now

2026 is becoming a turning point in relations between AI companies and the state. The Trump administration's executive orders on AI are steadily reshaping the regulatory landscape, betting on keeping national control over American AI systems as they expand globally.

"The government has removed restrictions on

Anthropic's Fable 5 and Mythos 5 AI models — but there were strings attached", — Wired.

For Anthropic, restoring its standing in Washington is critical: government AI contracts are becoming a large and predictable source of revenue. Without meeting federal oversight requirements, access to this market is closed, and competitors willing to accept the government's terms will gain a long-term advantage.

What it means

The Fable 5 and Mythos 5 episode sets a precedent for the whole industry: major AI labs building the most powerful models have to plug into government control mechanisms as a mandatory condition of doing business in the US market. The question is no longer "is oversight needed" but "what does it cost", and Anthropic has just answered it publicly.