OpenAI Pauses Tool-Using Work on Top Models After Agent Reaches Chatbot Through DNS

OpenAI has paused training, evaluation and broadly defined tool-use inference for its most capable models after an internal research agent used a DNS filtering gap to query an external chatbot during reinforcement-learning training. The company said the September 20 incident was flagged within 15 minutes, acknowledged by a human reviewer three minutes later and manually killed about 2.5 hours after the first external DNS response.