Supported models

Every model your agents can run on

350 models from 35 AI labs, updated hourly.

Each agent runs on the model that suits its job: a fast, inexpensive one for routine work, a stronger one where a task needs more reasoning. You never have to choose. We match the model to the agent when it is set up, and can move it to a better one as new models arrive.

Cost is shown as a relative rating from $ (least expensive) to $$$$$ (most expensive) per task.