OpenAI launches GPT-5.5-Cyber for vetted cybersecurity defenders

OpenAI released GPT-5.5-Cyber via its Trusted Access for Cyber program, a model variant with relaxed content filters purpose-built for penetration testing and vulnerability analysis. Access is restricted to vetted critical-infrastructure defenders, setting a precedent for tiered, use-case-specific model distribution in sensitive domains.