This article examines the logic underlying different models for assessing the college readiness of high school students. It focuses on benchmark scores that purportedly identify students who are college ready and presents the challenges of using threshold scores from a single assessment instrument to represent readiness. As well as providing logical and methodological reasons to be wary of threshold scores, analyses of Minnesota and national data illustrate their shortcomings. The article argues that assessments of college readiness should (a) use benchmarks with meaning and consequences for students, (b) employ multiple measures to provide readiness information more precise than from a threshold score derived from any single assessment, and (c) present readiness in terms of probabilities or likelihoods rather than as ready or not. (Contains 6 notes and 7 figures.)