U.S. Clears Anthropic's Claude Mythos 5 for Limited Release to 100+ Institutions

U.S. Lifts Block on Claude Mythos 5, Clearing Anthropic to Reach Over 100 Trusted Institutions

The U.S. government on Friday, June 27, 2026, lifted its two-week block on Anthropic's Claude Mythos 5, clearing the company to release its most powerful cybersecurity AI model to more than 100 U.S. institutions, including major corporations and government agencies. The decision, conveyed in a letter from Commerce Secretary Howard Lutnick to Anthropic's chief compute officer Tom Brown, follows an intense period of daily negotiations between the administration and the company — and signals the beginning of a new era of government oversight over frontier AI models.

The partial restoration of access to Claude Mythos 5 ends a shutdown that began on June 12, 2026, when the Trump administration imposed export controls on the model, forcing Anthropic to cut off all customer access — including foreign nationals working at Anthropic itself. The move had far-reaching consequences across the private sector, government agencies, and Anthropic's own workforce. Now, with Lutnick's letter in hand, Anthropic can begin reconnecting the model to vetted organizations focused on defending critical infrastructure.

What the Lutnick Letter Says — and What It Doesn't

The Commerce Secretary's letter is specific in its scope. It states that Commerce Secretary Howard Lutnick has "determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model" and that Anthropic has "committed to work with the U.S. government on protocols and standards and releases." Crucially, the letter specifies that "a license will no longer be required to export, reexport, or in-country transfer the Claude Mythos 5 Model to entities identified in Annex A to this letter and their foreign national employees, or to Anthropic's foreign national employees."

That last clause is significant. One of the most disruptive elements of the original export control directive was its blanket restriction on access for all foreign nationals — including Anthropic's own international staff. The new arrangement reverses that, allowing non-American employees at authorized organizations, as well as Anthropic's own non-American employees, to access Mythos 5 once again.

However, the letter is notably silent on Fable 5, the consumer-facing variant of Mythos that was also pulled on June 12. According to Semafor, people close to the talks say discussions around Fable 5 are moving forward, but no timeline has been confirmed. Fable 5's path to restoration remains uncertain.

Anthropic confirmed the development in an official post on X: "Today, the government notified us that Mythos 5, our strongest cybersecurity model, can be redeployed to a set of US organizations that operate and defend critical infrastructure."

A Commerce Department spokesman described the outcome as a product of rapid, focused diplomacy. "In just two weeks, we have worked diligently to ensure America remains the global leader in AI while safeguarding our security," said Benno Kass, the department's spokesman, according to Semafor.

How the Shutdown Began: Jailbreaks, China Concerns, and a Disputed Refusal

The sequence of events leading to the June 12 shutdown is tangled and, in parts, disputed. According to reporting from 9to5Mac, concerns reached the highest levels of the administration after Amazon CEO Andy Jassy told U.S. Treasury Secretary Scott Bessent about a purported jailbreak that left Anthropic's models vulnerable to misuse. Around the same time, the White House was already concerned about a separate but related issue: according to Semafor's June 13 reporting, the administration imposed export controls on Mythos partly over suspicions that a China-linked group had accessed the model.

The situation was further inflamed by an allegation from Trump adviser David Sacks, who claimed that when the Trump administration notified Anthropic of the jailbreak vulnerability, Anthropic co-founder and CEO Dario Amodei said the jailbreak was not a serious risk and refused to fix it. Anthropic has not publicly affirmed this characterization.

According to Engadget, the government's export control directive required Anthropic to restrict access not just for international customers, but for all foreign nationals — a sweeping requirement that resulted in the company blocking all customer access to both Mythos 5 and Fable 5 on June 12. The impact extended into the national security community as well: according to an implicator.ai report from June 24, 2026, parts of the National Security Agency lost access to Mythos 5 after the directive, as NSA analysts had been testing the cybersecurity model inside the agency.

Project Glasswing and the Model's Cybersecurity Track Record

Claude Mythos 5 had been available before the shutdown to a subset of organizations involved in Anthropic's Project Glasswing — a controlled cybersecurity consortium launched in April 2026. According to NBC News, early participants included infrastructure providers like Cisco and financial institutions like JPMorgan Chase. Before the export controls were imposed, Project Glasswing's initial cohort of roughly 50 partners had used Mythos Preview to identify more than 10,000 high- or critical-severity security flaws, according to Anthropic's official website.

On June 2, 2026, Anthropic expanded Project Glasswing to approximately 150 additional organizations in more than 15 countries, bringing total membership to around 200 organizations, according to CNBC. Major partners included Apple, Nvidia, Microsoft, CrowdStrike, and Palo Alto Networks. That expansion came just ten days before the government intervened with export controls — a jarring reversal for organizations that had just received access.

The restoration announced June 27 is more limited in scope than Project Glasswing's pre-shutdown reach. Access is now restricted to more than 100 U.S. institutions specifically identified in Annex A of the Lutnick letter, with the stated purpose of helping those organizations "operate and defend critical infrastructure," according to NBC News.

A New Precedent for AI Oversight — and OpenAI Is Already in It

The Lutnick letter's implications extend well beyond Anthropic. According to Semafor, it "marks the beginnings of a new regulatory regime that gives the US government control over the release of frontier AI models." That framing is supported by a parallel development on the same day: OpenAI also released its latest model, GPT-5.6, to a short list of government-approved partners.

According to CNN, the White House requested that OpenAI limit the release of GPT-5.6 to a small number of government-approved partners because of its advanced capabilities, with both OpenAI and the administration viewing GPT-5.6 as "on par" with Mythos. According to Axios, GPT-5.6 is available as a limited preview to around 20 companies whose participation has been approved by the government — a considerably narrower circle than the 100-plus institutions now cleared for Mythos 5.

OpenAI voiced measured concern about the arrangement in an official statement, saying the company does not "believe this kind of government access process should become the long-term default" and warning that it "keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them."

The regulatory backdrop also includes a broader executive action. President Donald Trump signed an executive order earlier in June 2026 asking AI companies with advanced models to voluntarily submit them for government review 30 days before release. According to CNN, however, the formal framework for that process has not yet been established, leaving companies navigating significant uncertainty about what future compliance will look like.

Anthropic's Financial Position Heading Into This Moment

The controversy arrives at a pivotal moment for Anthropic's business trajectory. According to Investing.com, the company filed a confidential draft S-1 with the Securities and Exchange Commission on June 1, 2026, and raised $65 billion in a Series H funding round led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital — setting a post-money valuation of $965 billion. The company's annualized revenue run rate crossed $47 billion in May 2026, up from roughly $10 billion the prior year, according to the same source.

That growth trajectory makes the two-week shutdown of Mythos 5 and Fable 5 a potentially costly disruption, both in direct revenue terms and in the reputational stakes of demonstrating that enterprise customers can rely on continued access to Anthropic's products. The partial restoration is a step toward stability — but with Fable 5's future still unresolved and an emerging regulatory framework still taking shape, the uncertainty is far from over.

What Comes Next

For now, Mythos 5 will flow to more than 100 U.S. institutions cleared under the Annex A framework, with the stated focus on critical infrastructure defense. Non-American employees at those organizations, along with Anthropic's own international staff, regain access under the new license terms.

Fable 5 remains in limbo. According to Semafor, people close to the negotiations believe the consumer model will eventually be released, but no timeline has been set, and the Lutnick letter does not address it.

The broader question — what role the U.S. government will play in approving, restricting, or shaping the release of powerful AI models going forward — remains wide open. The executive order calling for pre-release review exists, but the mechanism for implementing it does not yet. That gap will define the next phase of negotiations between frontier AI developers and federal regulators, with significant consequences for how quickly advanced tools reach the organizations and individuals who could use them.

For more tech news, visit our news section.

Why This Matters for Productivity and Personal Optimization

The push and pull between government oversight and AI access isn't just a policy story — it directly shapes which tools professionals, health systems, and enterprises can rely on to improve outcomes and protect the infrastructure underpinning daily life. As AI models become embedded in cybersecurity, healthcare, and productivity platforms, the stability and availability of those models becomes a personal and organizational concern. Staying informed about how access frameworks evolve is no longer optional for anyone building workflows around AI. Join the Moccet waitlist to stay ahead of the curve.