In addition to adapting to learner knowledge, the materials given to learners could be adapted to ensure they have the cognitive resources available for performing the learning task. In order to perform such adaptation, we need to be able to model the different domains of cognitive load (i.e., intrinsic, extraneous, and germane), which continues to be a logistical challenge. In this study, we examined whether item difficulty estimated using item response theory—a long-established and widely-used psychometrics method—could be used to model intrinsic load in adaptive learning environments. We found that item difficulty could be a viable option for modeling the intrinsic load of a task at the learning session level. We also found that learner experiences of intrinsic load varied according to their abilities. This finding underscores the need for modeling intrinsic load for creating adaptive systems that match task complexity with learner needs.