Anthropic Frontier Red Team Evaluates Cyber Capabilities Across AI Models
🤖 AI-generated content — The title and summary were produced automatically by artificial intelligence, without human editorial review.
In evaluations across 100 random tasks from an internal Binary Exploitation benchmark shared on September 29, 2026, Claude Mythos Preview achieved full control flow hijacks in 6% of trials, while GLM-5.3 succeeded in 4%. Earlier models like Claude Opus 4.6 and GLM-5.2 achieved zero successes.
Continue in the app — vote & join in ➔Source: Simon Willison (AI) · via ahirlevel.hu