OpenAI has introduced Patch the Planet, a project with security firm Trail of Bits that uses specialized AI models to automatically write, test, and merge security fixes for open-source software.
OpenAI has announced a new cybersecurity project called Patch the Planet to help protect open-source software from hackers. Created alongside security firm Trail of Bits, the initiative aims to fix real security flaws before automated AI hacking tools can find and exploit them.
The project addresses a growing problem for the open-source community: a massive wave of low-quality, AI-generated bug reports that has recently overwhelmed the independent developers who volunteer to maintain popular software. Instead of just pointing out errors, Patch the Planet uses AI to automatically build, test, and apply working security fixes. During a recent five-day trial, a team of 25 engineers used the system to quickly uncover hundreds of technical flaws and successfully push out dozens of official software patches.
To power these defense tools, OpenAI also introduced GPT-5.5-Cyber, a highly specialized security version of its AI. This model scored a record 85.6% on cybersecurity benchmarks, beating out rival systems from companies like Anthropic. Because the model is highly capable, OpenAI is keeping it strictly private, offering it only to verified government and institutional defenders through a restricted access program.
