GLM Model Scans Open-Source Projects, Uncovering 2,400+ Security Vulnerabilities

cedric_chee · x · 2026-08-14

The GLM model is being used to scan open-source projects and popular software at scale, with discovered vulnerabilities publicly disclosed on the Z.ai Security Disclosure Ledger.

The ledger currently contains 2,436 vulnerabilities, including 1,097 critical and high-severity flaws. These bugs have remarkably long lifespans, averaging 26.6 years before discovery, with the oldest dating back to 1981 (a 45-year span).

Recent high-severity disclosures affect core infrastructure components such as the Linux kernel, Apple Safari (WebKit), FreeBSD, and GStreamer.

Related event: Zhipu's GLM Model Scans Open-Source Projects, Uncovering Over 2,400 Vulnerabilities(2 posts)→

Original post →

More from Models

Models channel →