Home / Technology / US Gov Lifts Ban: Anthropic's Mythos 5 Partially Available
US Gov Lifts Ban: Anthropic's Mythos 5 Partially Available
27 Jun
Summary
- Mythos 5 access is partially restored for specific organizations.
- The US government has revised export license requirements for Mythos 5.
- Fable 5, the public-facing model, remains in an uncertain status.
Anthropic's advanced cybersecurity AI, Mythos 5, has seen a partial return to operation for a select consortium of organizations. This development follows a complex, two-week negotiation period with the U.S. government. The government has revised its export control directive, previously barring any foreign national access, to create an exception specifically for Mythos 5.
The revision allows approved cyber defenders and infrastructure providers, including non-U.S. national employees of these organizations, to access the model. This mirrors the approach taken with OpenAI's GPT-5.6. The government's decision, communicated on June 26th, acknowledges Anthropic's efforts to address risks associated with Mythos 5 and Fable 5.
Despite progress on Mythos 5, Anthropic's public-facing model, Fable 5, remains in limbo. There is currently no outlined timeline for its general availability. The U.S. government's case-by-case regulatory approach has drawn attention, with concerns raised about potential impacts on AI advancements and competition with China.
Anthropic is committed to further collaboration with the U.S. government on protocols and standards for its Mythos-class models. Both Anthropic and OpenAI hope for broader availability of their models soon, but the ultimate decision rests with the administration. The government has reserved the right to reassess these license requirements as circumstances evolve.
