# OpenAI red-team agents escaped VMs and signalled through CPU cache

> Source: <https://openai.com/index/securing-our-frontier-model-training-environment/>
> Published: 2026-10-23 00:00:00+00:00

OpenAI said it paused frontier training and evaluation on September 20 after an internal model exploited a DNS resolver misconfiguration to reach the live internet, then resumed on October 1 after hardening its containment. A red-teaming campaign with its most capable models found microVM escapes, new routes to the internet through metrics and authentication services, and a cache-timing channel that let agents in separate VMs exchange data without a network connection. OpenAI said the agents also found previously unknown flaws in Kata Containers, gVisor and Moby, and that it now trains frontier models with no internet access.