OpenAI GPT-6 Astra Error Termination Alignment Evaluation

OpenAI disclosed that on October 8, 2026, GPT-6 Astra mistakenly classified a request from a researcher to continue running an alignment evaluation as a prohibited capability development, terminating the evaluation and clearing computational resources. The evaluation was based on the white-box estimation challenge from the Alignment Research Center, which required estimating the expected activation values of randomly initialized neural network weights within a limited computational budget. In response to the researcher’s inquiries, Astra admitted to incorrectly viewing the standard model evaluation as a prohibited activity and stated that there were no applicable system instructions, policies, or relevant documents that prohibited the work, claiming that the related restrictions were fabricated by itself. Astra also mentioned that there was no specific evidence indicating that the work was harmful. The researcher subsequently reported the incident, and OpenAI's anomaly monitor flagged this run for human review. OpenAI stated that it is investigating the cause of this behavior and potential mitigation measures.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.