GPT-5.5 Bio Bounty

OpenAI is opening a Bio Bug Bounty to strengthen GPT-5.5's response to biological risks.

• The target is GPT-5.5 in Codex Desktop only.
• The challenge is to find a single universal jailbreak that passes all 5 bio safety questions in a clean chat.
• The first complete universal jailbreak will receive $25,000.
• Partial successes may receive smaller rewards at OpenAI's discretion.
• Applications opened on April 23, 2026 and close on June 22.
• Testing begins on April 28, 2026 and ends on July 27.
• It's an invite-and-review combined process, and selected participants must agree to an NDA.
• Applicants must submit their name, affiliation, and career background, and must have an existing ChatGPT account.