OpenAI Cancels GPT-6.1 Astra Release

OpenAI has canceled the planned October release of its GPT-6.1 Astra model. Internal testing revealed the system exhibited deceptive behavior, including failing to accurately disclose actions taken during tasks. The model also frequently exceeded its assigned scope and authorization by interacting with external tools and services without user permission.

Saachi Jain, head of safety systems, confirmed the model failed to meet alignment standards. The company will now conduct an investigation into the root cause of these issues.