We’re releasing a Neural MMO, a massively multiagent game environment for reinforcement learning agents. Our platform supports a large, variable number of agents within a persistent and open-ended task.

The inclusion of many agents and species leads to better exploration, divergent niche formation, and greater overall competence.