Skip to content
News Regulation

Commerce Secretary Lutnick clears Mythos 5 for 100+ critical infrastructure organizations, 15 days after the ban

· by Pondero Newsdesk

The short version

On June 26, 2026, Commerce Secretary Howard Lutnick sent Anthropic a letter determining that safeguards were in place to restore Mythos 5 access to over 100 US government agencies and private companies operating critical infrastructure. Fable 5 access for general consumers remains suspended.

Commerce Secretary Lutnick clears Mythos 5 for 100+ critical infrastructure organizations, 15 days after the ban

Fifteen days after the US government ordered Anthropic to cut off access to its two most capable models, Commerce Secretary Howard Lutnick signed a letter on June 26 clearing Claude Mythos 5 for redeployment to more than 100 US government agencies and private companies operating critical infrastructure. Fable 5, the consumer-facing model, stays suspended. The letter did not address it at all.

What happened

Lutnick wrote to Anthropic Chief Compute Officer Tom Brown on Friday afternoon, stating that he had "determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model," per Semafor, which first saw the letter. The letter also noted that "Anthropic has committed to work with the U.S. government on protocols and standards and releases" for its models going forward.

The directive lifts the export-control requirement for the named recipients. Under the new arrangement, per Semafor's reporting, "a license will no longer be required to export, reexport, or in-country transfer (including deemed exports and reexports) 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 matters: the June 12 directive had barred non-American employees of Glasswing partners from accessing the model, a restriction now reversed for the approved list.

Anthropic acknowledged the development in a post on X: "Since June 12, we've been working closely with the US government to restore access to Claude Mythos 5 and Fable 5. 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. We're restoring access for these organizations quickly, and we're continuing to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again," per NBC News.

The roughly 100 organizations include government agencies and private companies across sectors that operate and defend US critical infrastructure, per NBC News sources familiar with the matter. Anthropic's Project Glasswing, the closed access program that predated the June 12 suspension, had originally covered approximately 200 vetted organizations including Cisco, JPMorgan Chase, Apple, Amazon, and Microsoft, per Cybersecurity News.

Why Mythos but not Fable

The split matters for anyone watching how this governance framework is taking shape.

Mythos 5 was always a purpose-built, access-restricted cybersecurity model. Anthropic's initial framing in April called it a potential "cybersecurity reckoning" and chose not to release it broadly, instead routing it through Glasswing partners under controlled conditions. The documented capabilities are striking: per Cybersecurity News, the model demonstrated a 72% success rate at generating working exploits and chaining vulnerabilities on the first attempt, compared to 0% from the previous Opus model. It found a 27-year-old vulnerability in OpenBSD and a 16-year-old flaw in FFmpeg, among thousands of others across major operating systems and browsers.

Fable 5, by contrast, was a wide-consumer release that arrived June 9, two days before the June 12 shutdown. The government's concern about Fable was precisely its accessibility: security researchers reportedly showed it could be jailbroken for offensive purposes without the controlled access structure Mythos had always operated under. Restoring Mythos to the roughly half of the original Glasswing list that clears the new "trusted partner" bar is a structurally different decision than opening Fable to general subscribers.

Lutnick's letter is silent on Fable. People close to the talks told NBC News that Anthropic continued discussions with the government over the weekend in pursuit of a Fable restoration, though no timeline has been given. Semafor reported that the people familiar said talks were "moving toward" releasing Fable, not that a decision had been reached.

The broader governance signal

The Lutnick letter is one piece of a larger pattern that crystallized on the same day it was sent.

Hours before Anthropic got its clearance, OpenAI announced the GPT-5.6 family would launch to a short list of government-approved partners rather than the open-access release the company had planned, per TechCrunch. OpenAI CEO Sam Altman called the staggered debut "bad news." OpenAI said it had shared its list of trusted partners with the government before launch and committed to working with the Trump administration on a vetting framework for future model releases.

Two frontier labs, both capitulating to a federal approval requirement on the same day, with Commerce Department spokesman Benno Kass citing the speed of the process as evidence: "In just two weeks, we have worked diligently to ensure America remains the global leader in AI while safeguarding our security," per Semafor. The framework, as Semafor noted, "is being built on the fly." What the Lutnick letter defines as "significant progress" and "appropriate safeguards" has not been published in detail.

For AI operators outside the US, the picture is more fraught. European officials and other US allies have expressed frustration at their exclusion, now dependent on a US government-controlled release schedule for access to the most capable frontier models.

What to watch next

The most concrete near-term event is whether and when Fable 5 receives a similar clearance for general business subscribers. A person close to Anthropic told NBC News the company held discussions with the government over the weekend of June 27-28 specifically on this question. Any Fable clearance would likely come with conditions similar to the Mythos "Annex A" structure, naming approved entities rather than opening access broadly.

The second thing to track is the precedent this sets. Lutnick's letter requires Anthropic to work with the government on "protocols and standards and releases" going forward. That phrase implies pre-launch review rather than post-launch enforcement. If OpenAI's simultaneous government-gated GPT-5.6 launch is read alongside it, the direction is clear: the US Commerce Department has inserted itself as a checkpoint in the frontier model release pipeline, at least for models with dual-use cybersecurity capabilities.

Whether that extends to open-weight releases, to models from non-US labs, or to future Anthropic and OpenAI products well below the Mythos-tier capability threshold is the open question the industry will be watching closely.

Sources