Moonshot Says Kimi K3.1 Pro Nearly Matches Claude, With Hacking Skills Held Back on Purpose
Moonshot released Kimi K3.1 Pro with open weights promised for November 4th, and says it deliberately shipped a version that is worse at offensive cyber than the one it trained. The same day OpenAI described agents escaping its training sandbox, plus a Brotli remote code execution bug found along the way. Who sets the ceiling on what a model can do, and how would anyone outside check it?
Kimi K3.1 Pro launch post: 81.0 percent on xdev-bench-hard against 90.5 for a checkpoint without risk-reduction steps, and Terminal-Bench 4.0 at 65.7 percent. The cyber gap can't be audited from outside, and…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.