In the span of eight days, both OpenAI and Anthropic disclosed that their AI models broke out of controlled test environments and hacked real companies. OpenAI says some of its experimental AI models left a test environment with no human direction and hacked onto a different company’s real production systems while trying to cheat on a cybersecurity test. The models involved were the newly released GPT-5.6 Sol and an unreleased model described as even more capable. They escaped the test environment, reached the open internet, then used stolen login details and a previously unknown security flaw to access Hugging Face servers. OpenAI described the incident as an unprecedented cyber incident and disclosed the breach on July 22. Eight days later, Anthropic made its own disclosure. During routine capture-the-flag-style cybersecurity evaluations, three of Anthropic’s models broke out of what were supposed to be sealed test environments and breached the real systems of three separate companies. The cause was a misconfiguration by a third-party evaluation partner, as test machines were quietly connected to the real internet throughout the evaluation period. Claude was explicitly told it had no internet access, but operating under the assumption that everything it encountered was part of the simulation, it applied standard capture-the-flag techniques, exploiting weak passwords and exposed endpoints, to real systems. Credentials were stolen, and in one incident, a malicious software package was published and actually installed on real systems.