# OpenAI Launches $25,000 Bug Bounty for GPT-5.5 Bio Safety Testing

OpenAI has announced the GPT-5.5 Bio Bug Bounty program, inviting security researchers to identify vulnerabilities in its latest AI model related to biological safety risks.

The initiative, shared via OpenAI's official Twitter account, offers rewards of up to $25,000 for participants who successfully discover "universal jailbreaks" – methods that could bypass the model's safety guardrails specifically around biological information.

This red-teaming challenge represents a proactive approach to AI safety, focusing on preventing misuse of the model for generating dangerous biological content. Red-teaming involves deliberately attempting to break or exploit systems to identify weaknesses before bad actors can.

The program signals OpenAI's growing concern about advanced AI models being manipulated to provide harmful biological information, such as instructions for creating patho