GPT-5.6 and Claude Mythos 5 Face U.S. Oversight Delays

The Trump administration moved in late June 2026 to place federal oversight controls on two of the most capable AI models yet released, requiring OpenAI to limit its GPT-5.6 Sol rollout to select partners and restricting Anthropic’s Claude Mythos 5 to more than 100 vetted U.S. organizations. The interventions mark the most direct government involvement in commercial AI model releases to date.

How the Government Restricted OpenAI’s GPT-5.6 Release

OpenAI’s GPT-5.6 Sol will not receive a standard public launch. Instead of distributing the model broadly, OpenAI plans to share it only with a select group of close partners while the Trump administration reviews access requests. According to TechCrunch, CEO Sam Altman told staff at an internal meeting this week that the government would be “approving access customer by customer” during the preview period.

Altman reportedly added that if the limited release goes well, OpenAI hopes to follow with a broader general release “a couple of weeks later.” The agencies that requested the restricted rollout were the Office of the National Cyber Director and the Office of Science and Technology Policy, according to The Information, which first reported the story. OpenAI’s staffers also “worked closely” with the government on the upcoming release, The Information reported.

Anthropic’s Mythos 5 Gets Partial Clearance After June 12 Directive

U.S. Commerce Secretary Howard Lutnick sent a letter to Anthropic cofounder and chief compute officer Tom Brown on June 26, 2026, permitting access to Claude Mythos 5 for more than 100 U.S. organizations — including large corporations and federal agencies. The letter, obtained by WIRED, stated that Lutnick had “determined that appropriate safeguards are in place.”

The partial clearance follows a June 12 directive in which Lutnick’s office ordered Anthropic to disable access to both Mythos 5 and the consumer-facing Fable 5 model, citing national security authorities. “Anthropic has worked with the U.S. government to address risks associated with the Covered Models. These efforts have yielded significant progress,” Lutnick wrote in the June 26 letter.

An Anthropic spokesperson, Eduardo Maia Silva, told WIRED: “We received notice from the US government that Mythos 5, our strongest cybersecurity model, can be redeployed to a small group of cyber defenders and infrastructure providers. We are working to provision the approved set of providers and restore their access to Mythos 5 as quickly as possible.”

Fable 5 Remains Blocked as Talks Continue

The Commerce Department’s partial clearance does not extend to Claude Fable 5, the consumer-facing version of Mythos that Anthropic released with additional safeguards. According to WIRED, Anthropic is still in active discussions with the White House about restoring access to Fable 5, with talks expected to continue over the weekend of June 28–29, 2026.

A person familiar with the matter told WIRED that both parties are hopeful the resolution of this episode will help inform a lasting policy framework for future model releases. Maia Silva confirmed that Anthropic continues “to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again.”

The Broader Shift Toward Federal AI Model Review

These interventions reflect a policy direction that accelerated through June 2026. Earlier this month, President Trump signed an executive order directing certain AI companies to voluntarily submit new models to the government for testing and evaluation before public release, according to TechCrunch. That move came after the administration, which initially positioned itself as taking a hands-off approach to AI, began pushing for formal federal oversight of frontier models.

Anthropic had already moved in this direction voluntarily earlier in 2026, announcing that Claude Mythos would only be released to a limited group of partners through a program called Project Glasswing. The company argued the model was too powerful to release broadly without additional safeguards. The government’s June 12 directive, however, went further — effectively freezing even that limited rollout until federal review was complete.

What This Means

The events of late June 2026 establish a new de facto standard for frontier AI model releases in the United States: government review before broad deployment. Whether framed as voluntary cooperation or implicit pressure, both OpenAI and Anthropic are now operating under federal sign-off requirements for their most capable models.

The customer-by-customer approval process described for GPT-5.6 Sol is operationally significant. If the Office of the National Cyber Director and the Office of Science and Technology Policy maintain that review role beyond this initial release, it could add weeks or months to the timeline for any model deemed sufficiently capable. The partial clearance for Mythos 5 — covering more than 100 organizations but not general consumers — suggests the government is comfortable with controlled B2B deployment while remaining cautious about open access.

The unresolved status of Fable 5 is the most commercially consequential open question. As a consumer-facing product, its continued unavailability directly affects Anthropic’s revenue and user base. The outcome of weekend talks with the White House may set the template for how future consumer AI releases are handled.

FAQ

What is Claude Mythos 5?

Claude Mythos 5 is Anthropic’s most advanced AI model, described by the company as its “strongest cybersecurity model.” It was originally released under restricted access through a program called Project Glasswing before the U.S. government imposed a broader access freeze on June 12, 2026.

Why is GPT-5.6 Sol not getting a public release?

The Trump administration, through the Office of the National Cyber Director and the Office of Science and Technology Policy, asked OpenAI to limit the initial rollout of GPT-5.6 Sol to select partners while the government reviews access requests on a customer-by-customer basis, citing security concerns.

What is the difference between Claude Mythos 5 and Claude Fable 5?

Mythos 5 is Anthropic’s frontier cybersecurity-focused model, now cleared for deployment to more than 100 vetted U.S. organizations. Fable 5 is the consumer-facing version of the same underlying model, released with additional safeguards, and remains blocked pending ongoing White House negotiations as of June 26, 2026.