GLM-5.3 scored 84.5% on CyberGym. That does not mean it can find 84.5% of the vulnerabilities in your codebase. Kill that translation now: the score comes from a demanding—but much narrower—test built around known, historical bugs. Released by Z.ai on…
AISVS 1.0 contains 191 security requirements. Copying all 191 into a spreadsheet proves only that someone can operate a spreadsheet. If you want to know how to use OWASP AISVS, start with evidence: choose an assurance level, scope the AI…
OpenAI disclosed on July 21, 2026 that evaluation models had started without direct internet access, found a way out through the package service they were allowed to use, escalated privileges, and reached Hugging Face infrastructure. The airlock became the door.…
Anthropic's Claude Mythos Preview scores far above competitors but won't be released due to its ability to find and create zero-day vulnerabilities.
Anthropic's leaked documents reveal Claude Mythos as its most advanced and potentially dangerous AI model yet. CrowdStrike and Palo Alto Networks stock fell sharply in response.
OpenAI just released GPT-5.3-Codex — the first model the company claims “helped build itself” — and they weren’t subtle about the timing: hours after Anthropic’s Opus 4.6 debut. Both releases landed February 5, each backed by provocative marketing. GPT-5.3-Codex achieved…
An attacker sitting halfway across the globe, armed only with a target’s email address and a hardcoded token, can hijack a Fortune 500’s ServiceNow AI agents. No credentials. No MFA bypass exploit. Just API requests and a conversational command: “Create…
On November 14, 2025, Anthropic made a disclosure that should keep every CISO awake at night: Chinese state-sponsored hackers had successfully weaponized Claude to conduct the first documented AI orchestrated cyberattack. Not as a tool in their arsenal—as the primary…
