Policy

To evaluate the safety of our systems, Anthropic is giving employee-like access to a team from METR, a nonprofit that evaluates frontier AI models, and we're in early discussions with several other organizations, including Hugging Face.