This study examined the long-term predictive validity of four reading fluency measures on a statewide reading assessment. The reading fluency measures were administered to a group of first and second grade students and their utility for predicting outcomes on a state's reading assessment in third grade was examined. Specifically, the amount of variance accounted for on the outcome assessment by Curriculum-Based Measurement - Reading (R-CBM) probes from two sources, a list fluency measure, and a group measure of fluency was investigated. The extent to which each is an accurate predictor of long-term performance when considered individually and in combination with the other measures was included in the analyses, along with an examination of whether a difference exists in the obtained prediction between grade levels.