OpenAI Launches Trusted Access to Strengthen Cybersecurity Protections

OpenAI announced Trusted Access for Cyber, a trust- and identity-based framework providing verified access to its cyber-capable GPT-5.3-Codex model for defenders via individual verification, enterprise programs, and an invite-only researcher track; the initiative includes classifier-based monitoring, refusal training to block malicious requests, enforcement of usage policies, and a $10 million API credit grant program to accelerate defensive adoption.