OpenAI said on November 2, 2026 that it will run a new internal security sprint that pairs its teams with researchers from Cathedral Systems to test and strengthen its defenses using its own models.
Collaboration and testing approach
The company described the exercise as a way to bring “differentiated expertise and approaches” to finding weaknesses. Cathedral Systems researchers will work alongside OpenAI teams to challenge defenses, and Gavin Kliger, CEO & Co-founder, Cathedral Systems, was quoted as endorsing the collaboration and the use of models to continuously test systems.
OpenAI said it will use its most capable models to investigate findings, reproduce vulnerabilities in isolated environments and prepare patches for review. The previous sprint, the company reported, allocated 25% of product engineers to find and fix weaknesses and relied on Codex to support engineers’ investigations and patch preparation.
Rewards, resources and access
The announcement also covers program changes and resource commitments. OpenAI said it is improving its Security Bug Bounty program, “including more than quadrupling” rewards for critical findings to encourage external researchers to probe assumptions and defenses.
In addition, OpenAI committed $1 billion in subsidized access over six months for frontline defenders protecting essential services and digital infrastructure, and said it will make the same model tools available to Daybreak customers. The company plans to share architecture and lessons from the prior sprint, along with a blueprint intended to support cyber defenders across industry and government.
OpenAI stated it will continue to advance its models and expand trusted access while using them to strengthen its own systems. The announcement sets concrete measures: a new joint sprint with Cathedral Systems, enhanced bug bounty rewards, and a $1 billion, six-month subsidized access commitment announced on November 2, 2026.
Original source: OpenAI News