A multiple baseline design was used across probe sets to evaluate the effect of detect, practice, and repair (DPR) on the math fact fluency rates of a third-grade class. DPR was applied with a large group (n = 11) selected from an intact class (n = 17) in a manner that allowed each student to work on specific problems from a specific basic fact operation (i.e., subtraction, multiplication, or division) based on a daily assessment of their instructional needs. Prior to the implementation of DPR, students were computing an average of 18.4 digits correct per minute (DCPM). After using DPR for 11 sessions, students were computing an average of 31.2 DCPM. This 12.8 DCPM increase over 36 problems required approximately 132 min (12 min per session). Discussion focuses on the need for and challenges associated with developing intervention strategies and procedures that efficiently increase student skills at the class-wide level, with an emphasis on simultaneously meeting the needs of both high and low achievers by differentiating instruction.