According to The New York Times, a Chinese think tank representative recently approached Anthropic—a leading American artificial intelligence (AI) firm—seeking access to its newest AI model, Claude Mythos. The American company reportedly rejected the request on the spot.
Released in April 2026, Mythos is classified as “digital-weapons-grade” technology due to its advanced offensive and defensive cybersecurity capabilities. Currently, access is restricted to only 40 elite U.S. and UK institutions.
The incident triggered a high alert within the U.S. National Security Council. The timing is crucial as the Trump administration is currently finalizing an executive order on AI regulation while preparing a business delegation to China this week.
