Download Framework
OpenAI / Preparedness Framework v2
Framework for evaluating and managing catastrophic risks from frontier models. Defines High/Critical capability thresholds across tracked risk categories. Governed by Safety Advisory Group.
Type:
IndustryDomain:
Frontier AI Safety
Coverage:
Safety & Reputational HarmCybersecurity
Content:
0 Risks0 Controls
Version: 2025
Framework Definition
Risks and controls associated with the framework
Assessment Layer
Concrete evaluations linked to controls to assess pass or fail
No evaluation mapping defined yet.