OpenAI has opened a bio bug bounty for GPT-5.5, inviting researchers to identify biological misuse risks in the model. The program is described as part of OpenAI's efforts to address catastrophic biosecurity risks associated with advanced AI.
Rather than looking for generic alignment failures, the bounty targets harms involving biological weapons, pathogens, toxins, and related laboratory knowledge. The focus is on realistic uplift: whether GPT-5.5 gives someone with malicious intent meaningful help they could not easily get from ordinary sources.
Because this is a single announcement, there is no independent assessment of the program's effectiveness or scope. Still, the move signals that OpenAI is treating biosecurity as a concrete bug class to be found and fixed before it can cause harm.