OpenAI Launches Full-Scale Effort to Patch Open Source Bugs as It Takes on Anthropic’s Mythos (www.wired.com)

OpenAI has launched a comprehensive initiative called "Patch the Planet" aimed at enhancing the cybersecurity of open source projects amid escalating fears regarding AI hacking capabilities. Collaborating with the security firm Trail of Bits and vulnerability management companies, OpenAI's project will provide free security consulting services to open-source maintainers, helping them identify and fix vulnerabilities while integrating AI security tools into their workflows. This initiative is crucial as many open source projects struggle with limited resources to manage the increasing volume of security reports, exacerbated by AI-generated vulnerabilities. In conjunction with Patch the Planet, OpenAI announced an upgraded version of its security-specialized model GPT-5.5-Cyber, which scored 85.6% on the CyberGym benchmark, outperforming competitor Anthropic's Mythos 5. OpenAI is also releasing its Codex Security scanner as an app plugin. With funding and resources from OpenAI, Patch the Planet has already begun producing significant results, uncovering hundreds of bugs and developing numerous patches within its first week. This initiative not only aims to bolster the security of critical open source software but also educates maintainers on leveraging AI tools to improve their coding practices sustainably, setting a precedent for future AI-driven cybersecurity efforts.