Safety and content filtering models for policy enforcement.
Models can support multiple endpoints. We group them by model and list every supported API route.
