Push the frontier of
your model

Idler builds reinforcement-learning environments for model evaluation and training. Our environments are available off the shelf or customized to your needs, with a focus on coding, reasoning, and a broad spectrum of economically valuable knowledge work.