A standard API for reinforcement learning and a diverse set of reference environments (formerly Gym)
Gymnasium for multi-agent environments
Simple and easily configurable grid world environments for reinforcement learning