White House Blocks Anthropic’s Plan to Expand Access to Its Mythos AI Model

TLDR

• The White House is opposing Anthropic’s plan to expand its Mythos AI model to 70 more companies, bringing the total to around 120.
• Mythos was kept from public release due to cybersecurity risks and its ability to detect and exploit software vulnerabilities.
• A limited group of companies including Apple, Microsoft, and Nvidia already have access under a project called “Glasswing.”
• US officials are concerned Anthropic lacks enough computing power to serve more users without affecting government access.
• Unauthorized users accessed Mythos on the same day Anthropic first announced its limited release plan.

Anthropic unveiled Mythos in early April 2026, describing it as an AI model capable of detecting and exploiting vulnerabilities in critical software systems. The company said it could find security loopholes that had gone undetected for decades, even by human experts and automated tools.

Because of those capabilities, Anthropic decided not to release Mythos to the public. Instead, it launched a private program called “Glasswing,” sharing a version of the model with a small group of companies, including Apple, Microsoft, and Nvidia, to help improve their security infrastructure.

Anthropic then proposed expanding that access to roughly 70 additional organizations, which would bring the total number of companies using Mythos to around 120. That plan has now run into opposition from the White House.

White House Raises Security Concerns

A White House official told Bloomberg News that members of the Trump administration have informed Anthropic they do not support the expansion plan. The official spoke on condition of anonymity.

One key concern is computing capacity. US officials worry that sharing Mythos with more companies could reduce the government’s own ability to use the technology effectively, according to the Wall Street Journal.

A White House official told the Journal the administration is working to balance innovation and security while ensuring AI models are rolled out safely.

The opposition adds another layer of tension to an already strained relationship between the Trump administration and Anthropic. In February, President Trump ordered the US government to stop using Anthropic’s technology after Pentagon chief Pete Hegseth labeled the company a national security supply chain risk.

Anthropic is currently challenging those measures in court.

Recent Talks and Unauthorized Access

Relations between Anthropic and the White House showed some signs of improvement earlier in April. Chief Executive Dario Amodei met with US officials at the White House, and Anthropic described those talks as “productive.”

But the situation became more complicated when Bloomberg News reported that a small group of unauthorized users in a private online forum had gained access to Mythos on the same day Anthropic first announced its limited release plan. Anthropic confirmed it is investigating the incident.

Some critics have accused Anthropic of overstating Mythos’s capabilities. The company operates in a highly competitive space, with rival OpenAI developing similar technology.

Anthropic declined to comment on the White House’s opposition. The White House did not respond to a request for comment.

The post White House Blocks Anthropic’s Plan to Expand Access to Its Mythos AI Model appeared first on CoinCentral.