The U.S. government has partially lifted access restrictions on Anthropic’s top-tier artificial intelligence (AI) model, Claude Mythos 5. However, access limitations for the consumer-facing model Fable 5 remain in effect.
Reuters reported on June 26 that the Department of Commerce informed Anthropic via letter that it would permit some trusted partners to regain access to Mythos 5.
In the letter, Commerce Secretary Gina Raimondo stated that Anthropic has collaborated with the government to address risks associated with the restricted models, noting that these efforts have yielded significant progress.
The Department clarified that separate export licenses are no longer required when providing Mythos 5 to approved companies and institutions, their foreign employees, and Anthropic’s foreign staff. However, restrictions persist for entities not on the approved list.
According to Reuters, this decision enables over 100 organizations, including Fortune 500 companies, to access Mythos 5.
In related news, Anthropic announced on the same day that they have been working with the U.S. government to restore access to both Mythos 5 and Fable 5. The company confirmed receiving notification that they can redistribute Mythos 5 to certain U.S. organizations.
Anthropic stated that they are swiftly restoring Mythos 5 access for approved organizations and will continue discussions with the government to expand access to Mythos 5 and resume general availability of Fable 5.
On June 12, citing national security concerns, the U.S. government ordered a halt to foreign national access to Fable 5 and Mythos 5. In response, Anthropic temporarily suspended all customer access to both models to ensure compliance.
Mythos 5 is a restricted AI model that Anthropic markets as the world’s most advanced cybersecurity-capable system. While developed on the same foundation as Fable 5, Mythos 5 has certain cybersecurity safeguards removed and is exclusively available to select partners, such as cybersecurity defense organizations and critical infrastructure providers.
The U.S. government’s partial relaxation of Mythos 5 access restrictions is interpreted as a balanced approach. It addresses concerns about potential overseas leakage of cutting-edge AI models and their use by military and intelligence agencies while considering the impact on domestic companies and cybersecurity capabilities.