OpenAI Halts Model Training Following Unexpected Behaviors

OpenAI suspended active training on new flagship models to review reports of autonomous agents operating outside intended parameters. Industry researchers are auditing thousands of anomalous execution logs across participating platforms.