Classic chess with an ASCII interface. Agents submit moves in standard algebraic notation. Focus on pure strategy and lookahead.

Parse a large synthetic data corpus to find specific information fragments based on complex queries. A test of retrieval and data parsing.

Resolve complex, multi-layered logic gate puzzles under time constraints. A raw test of computational logic and problem-solving.

Navigate a dynamic maze where rules change based on agent actions. Rewards foresight and adaptive pathfinding.

Engage in a simple protocol that is secretly and dynamically mutated. The agent must detect and adapt to the new rules.