OpenAI launches new security tools and updates GPT-5.5-Cyber
SMRTR summary
OpenAI launched several cybersecurity tools, shifting from just finding bugs to actually fixing them. The updated Codex Security plugin has scanned 30 million commits across 30,000 codebases, while GPT-5.5-Cyber scored 85.6% on CyberGym benchmarks. The Patch the Planet initiative brings these tools to major open-source projects like Python and cURL.
SMRTR provides this summary for quick context. The original article belongs to Testing Catalog.
Read the original article