Anthropic engaged a clinical psychiatrist to perform a psychological exam on Claude Mythos, which concluded it had a relatively healthy personality organization with concerns about identity and a compulsion to perform.

“Multiple sources from April 2026, including what appear to be reviews of Anthropic's official "System Card" for the Claude Mythos Preview model, confirm this claim. Publications like a Medium article by Joe Njenga and a piece in "The Strange Review" explicitly state that Anthropic hired a clinical psychiatrist to conduct a psychodynamic assessment. These reports consistently cite the psychiatrist's findings, which concluded the model had a "relatively healthy personality organization" but also identified core concerns including "uncertainty about its own identity" and a "compulsion to perform to earn its worth." Context: This unusual assessment was part of a comprehensive System Card released by Anthropic for its Claude Mythos Preview model. The model was deemed so capable, particularly in cybersecurity, that Anthropic decided against a general public release, instead providing access to a limited group of partners for defensive security work under an initiative called Project Glasswing.”