[KongchangAI]
BenchmarkCyber Offense

Cyber Offense

A cyber offense and defense benchmark published by Wouse AI, in which AI agents attack a simulated enterprise network without human assistance. It is used to measure the attack capabilities of agents.

Timeline (last 90 days)

Nov 10

Wouse AI released the cyber offense/defense evaluation Cyber Offense, in which AI agents attack two simulated enterprise networks without human assistance.

Unverified50%
Nov 10

In the Cyber Offense evaluation, DeepSig V4.1 Flash scored 72.8%, second only to OpenAI's GPT-6Soft (80.8%).

Unverified50%

All Facts (2)

Source Articles