OpenAI has increased bug bounty rewards to up to $30,000 for critical flaws in its main products and up to $172,000 for exceptional findings when bonuses apply.
The company announced the changes on November 2 and said average awards at the midpoint of the ranges for medium-, high- and critical-severity reports had increased more than fourfold. The $30,000 maximum applies to critical flaws in products including ChatGPT, Codex and the API Platform. Exceptional critical findings can qualify for a base reward of up to $100,000.
Added incentives could raise the total to $172,000. New bonuses recognize reports that include useful demonstrations showing how a flaw could be exploited. OpenAI will reward researchers through a five-tier loyalty program.
It will also offer extra reward pools for short-term campaigns that test new features. OpenAI says clear evidence showing how a flaw can be exploited helps its teams assess the issue. It also helps them develop a fix.
The company is applying the updated reward ranges to eligible reports still awaiting awards. OpenAI says some researchers paid in the 30 days before the announcement may receive the difference if their findings qualify for a higher amount under the new structure.