OpenAI Lets Cyber Vendors Embed GPT-5.5 in Defenses

OpenAI is expanding access to its frontier models by allowing 30 cybersecurity vendors and service providers to integrate GPT-5.5 capabilities into their products and services.

See Also: AI Agents Introduce a New Insider Threat Model

The Daybreak Cyber Partner Program allows security vendors, consulting firms and managed service providers to integrate OpenAI’s cyber capabilities into products their customers already use. Vendors said they want to use OpenAI’s latest models to improve vulnerability prioritization, accelerate incident response, enrich threat intelligence, automate attack-path analysis and provide stronger context for security risks.

“Participating partners can use GPT-5.5 with Trusted Access for Cyber – our primary model for most defensive cybersecurity workflows – in the security products and services they provide to customers,” OpenAI wrote. “This allows their customers to benefit from the model’s defensive capabilities and make their software more resilient, but keeps direct model access in the hands of participating partners.”

OpenAI said its Daybreak cyber initiative is focused on helping defenders move beyond vulnerability discovery and toward automated remediation and patching at scale. By combining advanced AI models, security tooling, trusted industry partners and open-source initiatives, OpenAI aims to help defenders patch software faster and improve cyber resilience across enterprises and critical infrastructure.

“Models can navigate large codebases, reason through attack paths, validate hypotheses and surface security issues that might otherwise stay hidden,” OpenAI wrote Monday. “Defenders absolutely need access to these capabilities and also need tools to fix what we can now find – before attackers do.”

OpenAI’s push to allow cyber vendors to embed GPT-5.5 capabilities in their products comes less than three weeks after rival Anthropic extended access to Claude Mythos Preview to approximately 150 new organizations based in more than 15 countries. Anthropic’s update meant U.S.-based identity sector vendor Okta and data protection vendor Rubrik gained access to Mythos for the first time.

From Internal Testing to Embedding in Cyber Products

Up until now, OpenAI has restricted access to GPT 5.5 to vendors’ internal use. Putting GPT-5.5 inside customer-facing defenses is a first for the industry, and it means security vendors can now ship these cutting-edge capabilities to customers rather than just experimenting with AI behind closed doors. OpenAI in April unveiled a partner list for Trusted Access for Cyber concentrated in the financial services sector.

“We will also collaborate with program partners to continue to strengthen the safeguards, monitoring and abuse-prevention standards needed to deploy these capabilities responsibly across the security ecosystem,” OpenAI wrote Monday.

OpenAI Monday announced 19 cybersecurity product vendors and eight global systems integrators participating in Daybreak, many of which described how they intend to use GPT-5.5. NCC Group said it will research GPT-5.5 applications for cyber resilience and vulnerability discovery, while Cato Networks said it will explore AI-driven vulnerability discovery, prioritization and agentic defense capabilities.

Check Point Software said it plans to embed OpenAI cyber capabilities directly into customer-facing security products and services, while Darktrace said it will explore combining its behavioral AI models with OpenAI’s contextual reasoning capabilities to better connect technical events with business impact.

Proofpoint said it will use GPT-5.5 within managed security workflows to improve threat investigation, alert enrichment and incident response, while SpecterOps said will apply OpenAI capabilities to attack-path analysis, identity security and reverse engineering research.

Tenable said it will explore integrating OpenAI models into exposure management workflows to help organizations prioritize cyber risk, while TrendAI plans to integrate OpenAI capabilities into managed security services, vulnerability research and threat intelligence programs. OpenAI said it plans to continue expanding the Daybreak Cyber Partner Program to more organizations in the coming months.

“Security partners and practitioners can use our frontier models to strengthen their defensive tools and bring those capabilities to more organizations quickly,” OpenAI wrote. “The goal is to move beyond using models to find more vulnerabilities, towards a world of safer software and cyber resilience.”