AI Matters
Why it matters
Chinese AI model lags far behind US rivals on cyber exploit tests
The Decoder Kimi K3 trails frontier US models by a wide margin on cyber exploits, and distillation may explain why
UK and US safety testers found Moonshot AI's Kimi K3 scored 32 percent on an exploit benchmark versus 76 percent for leading US models, with weaker safeguards.
Why it matters
Highlights a real gap in AI safety testing as governments weigh how to handle increasingly capable foreign models.
Latest brief · Jul 24, 2026