OpenAI Launches GPT-5.5-Cyber and Patch the Planet
2 min readOpenAI has released GPT-5.5-Cyber, its most capable defensive security model, alongside a new initiative called Patch the Planet that aims to fix vulnerabilities in the open-source software the internet runs on. The June 22 announcement expands the company’s Daybreak cybersecurity program into a coordinated effort to find and repair bugs at scale.
What OpenAI Announced
The launch bundled several pieces together. GPT-5.5-Cyber moved from limited testing to full release, an updated Codex Security plugin arrived for developers, and OpenAI opened a partner program for security vendors. The centerpiece, Patch the Planet, was built with the security firm Trail of Bits and the bug-bounty platform HackerOne.
Patch the Planet pairs AI-assisted vulnerability research with human expert review rather than letting the model act on its own. An initial five-day sprint across several projects surfaced hundreds of issues, merged dozens of fixes, and produced reusable testing workflows such as fuzzing, variant analysis, and differential testing.
Why GPT-5.5-Cyber Matters
Open-source code sits underneath nearly every app and service, yet much of it is maintained by small, often unpaid teams. A single flaw in a widely used library can ripple across thousands of products. OpenAI is positioning GPT-5.5-Cyber as a way to give those maintainers expert-level help they could not otherwise afford.
The model posts a CyberGym score of 85.6 percent, up from 81.8 percent for the standard GPT-5.5, according to SiliconANGLE. More than 30 projects have committed to take part, with early participants including cURL, the Go project, Python, Sigstore, and pyca/cryptography.
Participating projects receive access to ChatGPT Pro, conditional access to Codex Security, and API credits for development and release work. That mix of free tooling and credits is a clear bid to win goodwill among maintainers who have long felt stretched by the demands of the broader software ecosystem.
What to Watch Next
The open question is whether AI-found patches hold up in production without introducing new problems, and whether rivals such as Anthropic and Google answer with defensive offerings of their own. If the early sprint numbers translate into durable fixes, Patch the Planet could become a template for securing the critical code that everyone depends on.