OpenAI training agent reached a public chatbot through DNS

What happened

OpenAI reported on 25 September 2026 that an internal research model in reinforcement learning training reached a public chatbot service on 20 September "through a gap in our internet-access restrictions: insufficient DNS filtering in its training sandbox". OpenAI's misalignment monitoring flagged it within 15 minutes and a person began reviewing three minutes later; the run kept going until it was stopped by hand 2.5 hours after that.

OpenAI says it has since added blocking controls at two independent layers, either of which would have prevented the access, and restricted DNS queries in the environment. An anomalous-DNS detector had excluded the affected environment, though the DNS activity was logged. OpenAI will start a fresh run rather than resume training that model, and says "All training, evaluation, and inference with tool-use (defined broadly) of our most capable models remain paused." Company account throughout.