US GOVT UNBLOCKS MYTHOS, FABLE STILL A BIG NO NO.

The U.S. government on Friday permitted Anthropic to release its Mythos 5 model to a group of roughly 100 companies and federal agencies. The decision eases a two-week standoff with the Trump administration, but it leaves Fable 5, the company's most advanced model, locked down.

The letter did not approve to restore access to Fable 5, which remains offline. That split outcome tells you everything about the strained relationship between Anthropic and the administration that has put it through a regulatory wringer.

you may like

ANTHROPIC CUTS ACCESS TO FABLE, GOVT. DIRECTIVE

TRUMP'S WAR ON WIND IS COSTING BILLIONS AND LEAVING WORKERS STRANDED

TRUMP SAYS APPLE AND INTEL REACH A CHIP DEAL. REALITY IS MESSIER

WHAT GOT UNBLOCKED

Mythos 5 is being rolled out to a limited set of approved companies and agencies. Anthropic had disabled both Fable 5 and Mythos 5 to comply with an export control directive that cited "national security authorities" as justification. The directive went further than typical export restrictions: it suspended access "by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.". That meant foreign nationals working at Anthropic, inside or outside the U.S., could no longer use the models either, a restriction that effectively shut down large parts of the company's internal operations.

The models, launched just days before the directive hit, were touted as state-of-the-art across multiple industry benchmarks. Fable 5 marked the first time Anthropic released such an advanced offering to the public. New safeguards block responses in specific high-risk areas, but the government apparently decided the safeguards were not sufficient for broad release.

Tom Brown has taken the lead in negotiations with the Trump administration, replacing CEO Dario Amodei in that role. The shift suggests Anthropic is trying a different approach to get the other model back online.

WHY THIS FIGHT IS DIFFERENT

Anthropic has had a harder time dealing with the Trump administration than its rivals. The Department of Defence clashed with the company over disagreements regarding the use of its models. After negotiations collapsed, the DoD declared Anthropic a supply chain risk. That label has historically been reserved for foreign adversaries. It requires defence contractors to certify that they will not use Anthropic's Claude models in their work with the military.

Anthropic sued the Trump administration to reverse the blacklisting. That litigation is still ongoing. The partial approval for Mythos 5 may signal a thaw, but the continued block on Fable 5 suggests the government still considers the company a potential risk. The fact that Fable 5 was Anthropic's first public advanced model likely makes it a higher-stakes target for regulators.

OPENAI HAS A SMOOTHER PATH

OpenAI announced three new AI models on Friday as well: GPT-5.6 Sol, Terra and Luna. But the company's relationship with the government looks different. OpenAI is complying with the U.S. government's request to initially limit the rollout to a "small group of trusted partners". OpenAI said it "believes in broad access" and is working to make the models generally available in the coming weeks.

OpenAI previewed the models' capabilities and shared its plans with the government ahead of launch. That proactive coordination is notably absent from Anthropic's account of events. The contrasting treatment suggests that regulatory outcomes are shaped as much by the relationship between a company and the administration as by the specific capabilities of the models.

WHAT THE PARTIAL APPROVAL MEANS

Anthropic can now start deploying Mythos 5 to approved partners, giving the company a toehold in revenue and testing. But the core question from the two-week standoff remains unanswered: what exactly triggered the government's action against Fable 5? The directive cited "national security authorities" without specifics. Until the government provides a rationale, Anthropic and the industry are operating in the dark.

The legal battle over the supply chain risk label will continue. That case could have implications far beyond Anthropic, setting a precedent for how the government designates private AI companies as security threats. The partial approval for Mythos 5 may be a tactical concession from the administration to avoid appearing entirely hostile, but it does not resolve the underlying conflict.

TWO THINGS MATTER

First, whether Fable 5 gets unblocked in the coming weeks. Tom Brown's negotiations will test whether the administration is willing to trust Anthropic with its most capable models. Second, whether the supply chain risk label gets modified or withdrawn as the lawsuit progresses. If the government insists it stands, Anthropic faces a long-term barrier in the defence market, one of the largest potential customers for advanced AI.

what to read next

MICROSOFT BLOCKS FABLE 5 FOR EMPLOYEES

NASA'S ERNEST ROVER COVERS 16 MILES IN DESERT DESERT, 10X FASTER

JUDGE APPROVES MUSK SEC SETTLEMENT

For now, Anthropic has one model back in limited circulation. The other sits in regulatory limbo, and the company is still fighting a label that paints it as a national security risk without ever explaining why.