Anthropic will restore global access to its most advanced AI models after the US government lifted restrictions imposed over cybersecurity and national security risks.
The move clears the way for the broader release of Claude Fable 5 and Claude Mythos 5, which had been placed under export controls while the company worked with federal officials to strengthen safeguards against misuse.
The restrictions were introduced on June 12 after US officials identified vulnerabilities they believed could be exploited for malicious cyber activities.
Anthropic subsequently worked with the US Department of Commerce to address those concerns before the controls were lifted.
According to Politico, Commerce Secretary Howard Lutnick informed the company that it had taken sufficient steps, in coordination with the US government, to mitigate the risks posed by the two frontier AI models.
Tightening Oversight of Advanced AI
Anthropic’s case comes as President Donald Trump’s administration expands its oversight of advanced AI releases.
OpenAI also limited the rollout of its GPT-5.6 model, making it available only to a select group of approved partners after complying with US government requests to restrict broader distribution.
Anthropic has proposed a standardized framework for evaluating AI jailbreaks, partnering with Amazon, Microsoft, Google, and other Glasswing participants to develop a shared approach to assessing security risks.
“This problem will become more acute in the coming months, as more models with powerful cybersecurity (and other) capabilities are trained, assessed, and released,” the company wrote in a blog post.
