A recent study by the British AI Security Institute and the U.S. Center for AI Standards and Innovation found that Moonshot AI's Kimi K3 performed poorly in offensive cyber tasks, scoring 32% compared to 76% for leading U.S. models. The results suggest that Kimi K3's safeguards are ineffective in blocking exploit development and simulated attacks. The disparity between Kimi K3's strong general benchmark scores and weaker cyber performance has raised concerns that the model may have been distilled from another AI, specifically Anthropic's models. This raises questions about the model's reliability and trustworthiness in high-stakes applications.