OpenAI Boosts Security Bug Bounty Payouts and Adds Loyalty Bonuses
OpenAI has significantly raised rewards for its Security Bug Bounty program, introducing new loyalty tiers, quality multipliers, and priority campaign bonuses.
OpenAI has announced substantial upgrades to its Security Bug Bounty program, increasing financial rewards and introducing new structures designed to encourage independent security researchers. By expanding compensation ranges and adding targeted bonuses, the organization aims to strengthen collaborative defense mechanisms across its software ecosystem.
Independent investigators play a vital role in identifying security flaws that internal testing might overlook. These updates sit alongside internal initiatives like the Defense Factory work, where engineers utilize AI models to detect weaknesses, replicate them, and evaluate patches. The latest program changes intend to foster an inclusive, community-driven security environment.
Higher Payouts and New Bonus Multipliers
Under the updated structure, the midpoint for medium, high, and critical findings has increased by more than four times on average. Core products like ChatGPT, Codex, and the API Platform now carry payouts of up to $30,000 for critical vulnerabilities. Exceptional critical discoveries remain eligible for up to $100,000, which can stretch up to $172,000 when combined with newly introduced bonuses.
To reward thorough investigation, OpenAI has rolled out a quality multiplier for reports that include a reliable proof of concept isolating the bug, detailing exploitation conditions, and demonstrating potential impact. Additionally, a five-tier loyalty program rewards researchers who maintain long-term partnerships, offering scaling bonuses to frequent contributors. Short-term priority campaigns will also feature dedicated bonus pools to encourage early testing on newly launched product features.
OpenAI is also applying these updates retroactively. Submissions that have not yet received a bounty will be evaluated under the new reward tiers. Furthermore, findings awarded within the 30 days prior to the announcement will receive a retroactive adjustment if the new guidelines yield a higher total.
What it means for developers
Security researchers and developers working with large language models can leverage these updated programs to secure higher compensation for finding vulnerabilities. Building secure applications is essential, especially when integrating multiple models into production environments. For those looking to experiment with integration, developers can try top AI models cheaply through one API at https://apixoai.online.
In-Person Collaboration in Oakland
To connect directly with the research community, OpenAI's security team will attend the PROMPT2PWN event in Oakland on November 20–21. The gathering provides an opportunity for security professionals to test designated targets in person and discuss collaborative vulnerability research.
Source: Rewarding defenders with bigger bug bounties — OpenAI News. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
