Lates News

Model Evaluation and Threat Research: Since completing the exercise, OpenAI said that all vulnerabilities we found affecting its production safety and security systems have been fixed.