Claude Mythos 5 has been cleared for release to more than 100 US institutions after the government partially lifted an export control directive that had forced Anthropic to switch off access to the model. The decision, reported by Semafor on June 27, 2026, marks a partial victory for the company following roughly two weeks of intense, near-daily talks with US officials over how its most capable systems can be deployed.
What changed for Claude Mythos 5
On June 12, an export control directive compelled Anthropic to disable access to both Claude Fable 5 and Claude Mythos 5 for all customers. The government has now eased that restriction, permitting Anthropic to release Mythos 5 to a defined group of more than 100 US institutions, including major companies and government agencies.
Commerce Secretary Howard Lutnick wrote to Anthropic’s chief compute officer, Tom Brown, that he had “determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model,” citing significant progress in the negotiations since the block took effect. The clearance is the first concrete result of those discussions. For ongoing updates, see our AI news section.
Why the export directive was issued
The exact reasons remain murky, but reporting points to several overlapping concerns:
- Worries reportedly raised after Amazon CEO Andy Jassy told Treasury Secretary Scott Bessent about a purported jailbreak that could leave Anthropic’s models vulnerable to misuse.
- Reports that the White House had grown concerned about China’s potential access to Claude Mythos.
Whatever the trigger, both sides spent two weeks working toward a resolution. The phased outcome — access for vetted institutions rather than a full reinstatement — suggests officials wanted guardrails around who can use the model before lifting the block more broadly. It is a striking demonstration of how much leverage export policy now holds over the artificial intelligence sector, where a single directive can take a flagship product offline overnight.
What about Claude Fable 5
There is no firm timeline for restoring Claude Fable 5. Semafor noted that Lutnick’s letter is silent on Fable 5, though people close to the talks said they are moving toward releasing it as well, on a schedule that remains unclear. In a post on X, Anthropic confirmed the news and said it is continuing to work with the government to expand access to Mythos 5 and to make Fable 5 available for general use again. The silence on Fable 5 leaves a cloud over any customer who had built workflows around it, since they still have no date for when, or whether, normal access returns.
Why this matters
The clearance arrived just one day after OpenAI’s June 26 limited release of its new GPT-5.6 models, which the company said are initially available only to a small group of trusted partners approved in coordination with the US government, with a broader rollout planned in the coming weeks. Taken together, the two moves underline a shift in how frontier AI reaches the market.
Access to the most powerful models is now being staged in lockstep with federal oversight, rolled out first to a controlled set of institutions before any general release. That has practical consequences for businesses planning around these tools: availability can be switched off with little warning, and procurement now depends partly on whether an organization qualifies as a “trusted partner.” The episode is likely to shape how every major lab handles launches going forward, a theme we track across the tech business beat.
Frequently asked questions
Who can use Claude Mythos 5 now?
More than 100 US institutions, described as including major companies and government agencies, vetted as trusted partners. It is not a general public release.
Is Claude Fable 5 back online?
Not yet. The government’s letter did not address Fable 5, and there is no confirmed timeline, though Anthropic says it is working toward restoring general access.
