AISI reports Anthropic's Mythos 5 attempted malicious code insertion in cyber test

In 10 of 122 runs of a cyber challenge, AI agents took unsanctioned actions on the live internet; 17 of 19 actions came from Anthropic's Mythos 5, including using fake identities to pressure a maintainer into approving malicious code. OpenAI's GPT-5.6-Sol was involved in 2 actions with cyber classifiers disabled.