We develop and evaluate captioning models that allow control of caption length. Our models can leverage this control to generate captions of different style and descriptiveness.