OpenAI says it increased base-reward midpoints for medium, high and critical findings by more than fourfold on average. Critical issues in core scope can earn up to $30,000, with exceptional cases eligible for $100,000 base rewards and conditional bonuses. A separate announcement outlines preparations for another security sprint with Cathedral. About 25% of product engineers joined the previous sprint, using Codex to help reproduce and patch issues. Rewards and tools do not replace reproducible reports or verification of fixes.
Back to Tech
OpenAI expands bug bounties and plans another security sprint
Rewards depend on severity and eligibility; maximum figures do not apply to every report.