OpenAI expands bio bug bounty program

OpenAI has expanded its GPT5.5 Bio Bug Bounty into an ongoing private program called the OpenAI Bio Bounty Program. The effort is focused on finding universal jailbreaks that can defeat the company’s biosafety challenge in frontier models, beginning with GPT5.6 and continuing with future versions. The company said the top reward for a universal jailbreak has been raised from $25,000 to $50,000 for both GPT5.6 and GPT5.5. OpenAI said smaller awards may also be given for partial results at its discretion. Testing for the original GPT5.5 program will continue until July 27, 2026. After that date, only GPT5.6 will remain in scope. Researchers interested in joining must apply through a rolling application process, and accepted participants will be onboarded to the platform and required to sign an NDA.