OpenAI Launches GPT-5.5 Biosafety Vulnerability Bounty Program, Offering $25,000 Reward

OpenAI has launched a biosafety vulnerability bounty program for GPT-5.5, inviting researchers to find universal jailbreak methods that can bypass five biosafety challenges. Those who succeed can get a reward of up to $25,000.

GPT-5.5 is OpenAI's most intelligent and intuitive AI model to date, performing excellently in Agentic Coding, computer usage, and scientific research fields. Compared to GPT-5.4, it consumes fewer tokens when completing the same tasks.