OpenAI Pulls Special Access From Security Researchers Just Over a Week After Launching It

OpenAI removed special access for at least five security researchers on August 19, 2026, nine days after launching its Daybreak Blue tier. That tier lets approved researchers use advanced AI models with fewer safety limits for cybersecurity defense work. OpenAI blamed a "technical issue" and told affected users to reapply and repeat the verification process.