OpenAI raises bug bounty maximum to roughly $172,000

OpenAI has increased payments for security researchers who identify vulnerabilities in its products. Critical flaws in core services such as ChatGPT and Codex can now qualify for rewards of up to $30,000. Exceptional findings may bring payments of up to $100,000, while quality, loyalty and campaign bonuses can raise the total to roughly $172,000.

The revised program includes a five-tier loyalty scheme and temporary campaigns offering higher rewards when new features are launched. Researchers who reported a vulnerability during the previous 30 days will receive the difference between the former and current payment rates. The new maximum places OpenAI’s general rewards program near Google’s, although it remains below the highest payments offered by Apple and Microsoft for operating-system exploits.

OpenAI is also planning a security sprint with the firm Cathedral. Its specialists will work with OpenAI engineers to test the company’s systems for weaknesses. OpenAI said that during the previous sprint, 25 percent of its product engineers took part in vulnerability-hunting work.