Anthropic’s most advanced cybersecurity model, Mythos 5, has been partially restored following a two-week regulatory standoff with the Trump administration. On June 26, US Commerce Secretary Howard Lutnick notified Anthropic co-founder Tom Brown that revised license requirements would permit a carefully vetted group of domestic cyber defenders and infrastructure providers to regain access to the model. The decision follows Anthropic’s collaboration with federal authorities to establish new risk-mitigation protocols for both its Mythos and Fable model families. The restoration operates under a limited-exception framework rather than a full lifting of the June 12 export control directive, which previously barred all foreign nationals, including Anthropic’s international staff, from accessing either model. Under the new terms, non-US nationals affiliated with approved organizations and Anthropic employees who are not US citizens may interact with Mythos 5. Anthropic spokesperson Danielle Ghiglieri confirmed the company is currently provisioning these approved providers and expressed confidence that continued government collaboration will eventually expand access and restore general availability for the public-facing Fable 5 model. The regulatory pause had drawn significant scrutiny from the US AI sector, with industry leaders warning that prolonged export restrictions could cede strategic advantages in AI-driven cybersecurity to foreign competitors, notably China. Domestic government agencies, including the National Security Agency, also suffered operational delays while their access to the model remained suspended. The resolution mirrors a recently announced arrangement for OpenAI’s GPT-5.6, establishing a parallel precedent for how top AI labs will navigate federal oversight. Both companies have advocated for a structured, repeatable licensing process rather than ad hoc government reviews. Despite the immediate relief for approved partners, the broader deployment timeline remains uncertain. Secretary Lutnick emphasized that all original conditions of the June 12 directive remain in force for unapproved users and reserved the right to adjust access parameters as technological and geopolitical conditions evolve. The compromise highlights an ongoing tension between national security oversight and commercial innovation, as AI developers push for streamlined regulatory pathways that balance risk management with rapid enterprise adoption. Anthropic and its competitors continue to negotiate with federal officials on a permanent cyber Executive Order framework that would standardize model releases and restore broader access in the coming months.