OpenAI has expanded its Daybreak Cyber Partner Program, bringing its frontier cybersecurity models into the services of a group of established security and technology providers, the company announced on 10 August 2026.
Partners named include Accenture, IBM, Capgemini, Cognizant, EY, KPMG, PwC, NCC Group and SpecterOps, alongside technology partners Palo Alto Networks, CrowdStrike, Cisco, Sophos, Akamai, Fortinet and Cloudflare. Depending on the engagement, partners can use the capabilities for vulnerability discovery and validation, red teaming, penetration testing, incident response and remediation.
Partners access either Daybreak Blue, aimed at broad defensive workflows, or Daybreak Red, reserved for more tightly governed work such as red teaming and penetration testing. OpenAI said access to the underlying models stays with the approved partner rather than passing directly to the end customer, with safeguards that can include identity verification, defined testing scopes, logging, monitoring and human oversight.
