We model the production of quantified referring expressions (QREs) that identify collections of visual items.To address this task, we propose a method of perceptual cost pruning, which consists of two steps: (1) determine what subset of quantity information can be perceived given a time limit t, and (2) apply a preference order based REG algorithm, such as the Incremental Algorithm (IA), to this reduced set of information.We demonstrate that this method successfully improves the human-likeness of the IA in the QRE generation task by successfully modeling humangenerated language in most cases.