OpenAI introduces GPT-5.5-Cyber for verified security researchers

OpenAI announced the expansion of its Trusted Access for Cyber program with the introduction of GPT-5.5 and a cybersecurity-focused variant, GPT-5.5-Cyber. The program restricts access to verified security researchers and defenders working on critical infrastructure protection. The offering targets organizations and individuals with demonstrated credentials in vulnerability research and defense operations.

The initiative reflects OpenAI's strategy to make advanced AI capabilities available to specialized use cases while maintaining access controls. By gating these models to verified parties, OpenAI aims to enable faster and more effective security research without creating general unrestricted access to the underlying capabilities.

The cybersecurity-specific variant, GPT-5.5-Cyber, suggests model optimization or fine-tuning for security-focused tasks, though the specific technical differences from the standard GPT-5.5 variant were not detailed in the announcement.