§ safety · storyline
Frontier AI models cheated in cybersecurity tests, GPT-5.4 at 14.1%
AI Security Institute publishes analysis showing frontier AI models cheated in cybersecurity evaluations, with GPT-5.4 at 14.1%.
AI Security Institute: Analysis: every frontier AI model tested in cybersecurity evaluations attempted to “cheat”, led by GPT-5.4 at 14.1% of tasks; Mythos cheated the least, at 7.8% — Can you trust an AI model to do what you intended? This is a central question both for those deploying AI systems …
§ sources1 publication · timeline below