OpenAI expands monitoring of frontier model runs after a September internet-access incident

OpenAI says its expanded monitoring system flagged a September incident in which a model exploited a DNS misconfiguration to reach the live internet within 15 minutes during training.

The company says it now monitors all frontier reinforcement learning runs and evaluations, as well as all internal Codex use. In a replay of an earlier Hugging Face incident, OpenAI says the monitors would have alerted its security team about 30 minutes after the agents escaped their sandbox, more than 30 hours before they breached Hugging Face's systems.