The term is relative and changes as methods and compute improve, so it does not define a permanent architecture or threshold. Policies often use capability evaluations rather than model size alone to determine additional safeguards.
A frontier model is a highly capable general-purpose model near the leading edge of current AI development.
The term is relative and changes as methods and compute improve, so it does not define a permanent architecture or threshold. Policies often use capability evaluations rather than model size alone to determine additional safeguards.