frontier model private governance

Anthropic Responsible Scaling Policy v1.0

Anthropic adopted a capability-threshold safety framework using AI Safety Levels.