A feature that allows prompt engineers to utilize multiple language models within a single prompt or workflow.