• June 12: Sudden shutdown over security fears
• Two‑week standoff and industry shock
• June 26–27: Mythos 5 partially restored
• Fable 5 still in limbo — but perhaps not for long

US Government Partially Lifts Restrictions on Anthropic’s Mythos 5 AI Model The U.S. government has begun unwinding an extraordinary clampdown on Anthropic’s latest AI systems, cautiously reopening access to the Mythos 5 cybersecurity model while keeping its more widely used sibling, Fable 5, largely offline.

§June 12: Sudden shutdown over security fears

On June 12, the Trump administration issued an export‑control order that forced Anthropic to disable both Mythos 5 and Fable 5, barring access by foreign nationals, including many of the company’s own employees. Officials were concerned that the models’ guardrails could be bypassed by security researchers, raising potential national security and cyber‑offense risks.

§Two‑week standoff and industry shock

The abrupt removal of the models “created uncertainty within the industry and among allies around the world whose cyber capabilities were stunted by the loss.” For developers, it was “unprecedented and deeply jarring — a top-tier model, already in users’ hands, pulled offline due to government intervention.” Some builders tried to adapt: one developer noted that while “we eagerly await Fable 5’s return,” they pushed alternative systems to extreme speeds in the browser.

§June 26–27: Mythos 5 partially restored

After two weeks of negotiations, Commerce Secretary Howard Lutnick informed Anthropic that “appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model,” revising license requirements for a limited carve‑out. Mythos 5 can now be used by more than 100 specified U.S. companies and agencies, including their foreign‑national staff and Anthropic’s own non‑U.S. employees. Anthropic said Mythos 5 “can be redeployed to a set of US organizations that operate and defend critical infrastructure” and that it is moving quickly to restore their access.

§Fable 5 still in limbo — but perhaps not for long

The carve‑out does not yet extend to Fable 5, the public‑facing Mythos‑class model, which “appears to still be in limbo, with no apparent timeline for a rollout agreement.” However, other agencies have already concluded Fable 5 “can safely return to the wild,” and insiders expect limits “could be lifted as soon as this coming week,” pending sign‑off from the Pentagon and NSA.

At the same time, OpenAI has been asked to restrict preview access to its GPT‑5.6 models to “a small group of trusted partners,” a process the company warned “should not become the long-term default,” arguing it keeps top tools from users and cyber defenders who need them.

The partial reopening of Mythos 5, combined with continued controls on Fable 5, underscores an emerging, ad‑hoc model‑by‑model regime in which powerful AI systems are treated more like sensitive dual‑use technologies than standard software.

Continue reading https://foxvector.com/stories/019f0a1d-bf4b-123f-707b-2e633b5e2fbe