Tag
A new open-source benchmark called Age of LLM tests LLM reasoning through a turn-based nuclear strategy game with fog of war, diplomacy, and bluffing, offering a more dynamic evaluation than traditional multiple-choice benchmarks.