June 5, 2026 Canada’s federal government now has access to Anthropic’s Mythos AI model, a cybersecurity system so powerful that the company has restricted access to a small group of trusted organizations. AI Minister Evan Solomon confirmed Tuesday that Canada has joined Project Glasswing, Anthropic’s initiative designed to help governments and major organizations identify security vulnerabilities before advanced AI hacking tools become more widely available.
Speaking to reporters after a cabinet meeting in Ottawa, Solomon said the Canadian government is participating in the program through the Canadian Centre for Cyber Security.
“I can confirm that the Canadian government is part now of Project Glasswing, which allows companies to have access to Mythos,” Solomon said.
He added that access to the model is an important step in helping Canada protect its institutions and critical systems from emerging cyber threats.
Anthropic announced earlier Tuesday that it was expanding Project Glasswing to an additional 150 organizations across more than 15 countries. While the company did not identify the new participants, it said the organizations span sectors including health care, power, water, communications, hardware, and government.
Solomon said he could not disclose which Canadian companies, if any, have access to the system.
Anthropic unveiled Mythos earlier this year, describing it as a model with unusually advanced cybersecurity capabilities. At launch, the company said the system was so capable at hacking and security-related tasks that it would only be made available to a limited group of trusted organizations.
The company created Project Glasswing to help secure critical software infrastructure against what it described as potentially severe risks to public safety, national security, and the economy if similar capabilities were misused in cyberattacks.
When Anthropic first announced the project, Solomon said restricting public access to Mythos was “the responsible path” because it gave defenders a head start before comparable tools became more broadly available.
According to Anthropic, organizations already using Mythos — including Apple, Google, Amazon, and Microsoft — have identified more than 10,000 security flaws that were classified as highly severe or critically severe.
The company has said Project Glasswing is intended to strengthen major software systems before Mythos-level capabilities become available from other AI providers. Anthropic estimates that could happen within six to 12 months. At the same time, the company has stated that it plans to make Mythos-class models available to all customers in the coming weeks with additional safeguards in place.
Canada’s participation in the program comes ahead of the federal government’s expected release of a new national AI strategy later this week.
Solomon said the strategy will focus heavily on building public trust in artificial intelligence while also addressing worker impacts, sovereign computing infrastructure, support for Canadian AI companies, international co-operation, and AI education and training.
He also said the government is working on new privacy and online safety legislation alongside Culture Minister Marc Miller and Justice Minister Sean Fraser. While Solomon would not say whether the legislation will be introduced before Parliament’s summer break, he said protecting Canadians, particularly children, privacy, and personal data remains a priority.
