Reinforcement-learning environments for training frontier machine-learning models. Aimed at researchers and ML engineering teams building and evaluating advanced RL agents.