Reports, citing Bloomberg, indicate that Anthropic’s Mythos model was accessed by unauthorized users on its launch day during an early access testing phase.
A company spokesperson confirmed that an internal investigation is underway to determine the circumstances of the breach, noting that the incident occurred while the model was being rolled out to a limited group of users.
The event highlights growing security challenges facing AI companies, particularly during early deployment stages when systems may be more vulnerable.
It also raises broader concerns about securing advanced AI infrastructure, especially as such models become increasingly integrated into critical domains like cybersecurity and strategic analysis.