OpenAI has increased its maximum bug bounty payouts to $172,000 as part of a major ramp-up in the company's security efforts. Under the updated program, critical vulnerabilities discovered in core products such as ChatGPT and Codex can now earn outside researchers up to $30,000 per finding. Exceptional security discoveries can yield up to $100,000, and when combined with quality, loyalty, and campaign bonuses, the total compensation can reach roughly $172,000.
The restructuring places OpenAI's rewards program in a similar tier to Google's vulnerability rewards, though it remains below the highest payouts offered by Apple and Microsoft for operating system exploits. The overhauled program introduces a five-tier loyalty structure alongside time-limited campaigns that offer boosted payouts tied directly to new feature launches. Additionally, OpenAI announced that anyone who reported a vulnerability within the thirty days prior to the update will retroactively receive the difference between the old and new rates.
Alongside the financial reward adjustments, OpenAI is organizing a new security sprint to further stress-test its systems. The security firm Cathedral will collaborate directly with OpenAI's internal engineers during this initiative. The company noted that a previous security sprint involved twenty-five percent of OpenAI's product engineers dedicating their time to identifying and addressing system vulnerabilities.
This security overhaul matters because it incentivizes external researchers to find and fix critical flaws in major AI products before malicious actors can exploit them.