AIThe Decoderabout 1 hour ago
Kimi K3 trails US models on cyber exploits
1 min read
Moonshot AI's Kimi K3 scored 32 percent on ExploitBench, far behind leading US models at 76 percent, in tests by the British AI Security Institute and the U.S. Center for AI Standards and Innovation. Its safeguards failed to block exploit development or simulated attacks. The gap between general benchmark scores and weak cyber performance fits allegations of distilling Anthropic's models.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
