US lifts block on Anthropic's Claude Mythos 5 for 100-plus institutions

The US government on Friday lifted its block on Anthropic's Claude Mythos 5, allowing the AI company to release its most powerful model to more than 100 institutions including major corporations and federal agencies.

"This marks a new chapter in government-industry cooperation on frontier AI safety," Dario Amodei, chief executive officer of Anthropic, said in a statement. "We've worked closely with regulators to ensure Mythos 5 meets the highest standards."

The clearance follows months of negotiations after the White House initially blocked the model's release in April over concerns about its capabilities. Anthropic agreed to implement tiered access controls and third-party auditing, according to a person familiar with the matter. The approved institutions span sectors including finance, health care, defense and energy.

The decision gives Anthropic a competitive edge over OpenAI, whose GPT-5.6 rollout remains constrained by similar government restrictions. The AI market is projected to reach $1.3 trillion in spending by 2032, and regulatory approval has become a key differentiator for enterprise adoption.

How Anthropic Won Back Washington's Trust

The White House had frozen Mythos 5's release in April after internal assessments flagged the model's advanced reasoning capabilities as a potential national security risk, according to two people familiar with the discussions. Anthropic responded by restructuring its safety protocols, including the addition of an independent oversight board and real-time monitoring systems that report to the US AI Safety Institute.

The company also agreed to limit initial deployment to US-based institutions with verified security clearances, a concession that helped rebuild trust with the administration. The approach contrasts with OpenAI's strategy, which has been more resistant to government-imposed deployment caps, according to a person familiar with both companies' regulatory interactions.

The Competitive Landscape Shifts

The approval positions Anthropic to capture enterprise clients that have been hesitant to adopt AI models without government validation. Microsoft, Amazon and Google have all expressed interest in integrating Mythos 5 into their cloud platforms, according to industry sources.

OpenAI, meanwhile, faces continued pressure. The Trump administration has pressed the company to limit GPT-5.6's release, with reports indicating the model will be available in three variants — Sol, Terra and Luna — each with different capability tiers subject to government oversight. The restrictions have slowed OpenAI's enterprise sales cycle, giving Anthropic an opening in the corporate market.

What's at Stake for Investors

Anthropic, valued at $61.5 billion in its most recent funding round, has not disclosed Mythos 5's training cost, but comparable frontier models typically require $4 billion to $6 billion in compute spending. The regulatory green light unlocks revenue from government contracts and enterprise subscriptions, which could accelerate the company's path to profitability.

For the broader AI sector, the decision signals that the US government is willing to clear advanced models when companies meet safety requirements — a framework that could set a precedent for future releases. Nvidia, whose H100 and B200 GPUs power most frontier AI training, stands to benefit from increased model deployment, while cloud providers competing for AI workloads face a more fragmented regulatory environment.

This article is for informational purposes only and does not constitute investment advice.