OpenAI announced a partnership with Hacktron, a security research team, to jointly find vulnerabilities in OpenAI‘s systems. The collaboration will focus on high-priority areas, especially how OpenAI‘s systems interact with third-party services. As part of the agreement, Hacktron is gaining expanded access to OpenAI‘s frontier models through Daybreak Red.

OpenAI says it is applying the same urgency to its own cyber defenses that it brings to accelerating cyber defense with frontier AI. Earlier in the week, OpenAI announced that Cathedral‘s researchers would join its next security sprint, and it raised its bug bounty rewards to incentivize independent researchers to find issues that internal testing might miss.

OpenAI became familiar with Hacktron through its Security Bug Bounty program and was impressed by the impact of the vulnerabilities Hacktron has discovered and disclosed. Hacktron builds AI agents for code review, penetration testing, and red teaming. Its research has uncovered vulnerabilities in how applications, cloud infrastructure, and identity systems interact. OpenAI notes that as AI agents connect services in new ways, this combination of skills is increasingly valuable, and working with Hacktron will help test those connections more deeply and strengthen defenses.