This study investigates the relative effects of two types of input modification—simplification and elaboration—on Korean high school students' EFL reading comprehension. Six English reading passages in one of three forms—(a) baseline, (b) simplified, or (c) elaborated—were presented to 180 students, who were divided into two proficiency levels (i.e., high proficiency and low proficiency). Comprehension was assessed by an 18-item multiple-choice test, which included items for assessing (a) general, (b) specific, and (c) inferential comprehension. In addition, students' perceptions of their comprehension were measured by their responses on a 6-point unipolar scale. The test data were analyzed by a 2-by-3 analysis of variance, with least significant difference tests used in post hoc analyses. The results support the suggestion that input should be modified in the direction of elaboration rather than by artificial simplification, because elaboration retains more nativelike qualities than and is at least equally successful as—if not more successful than—simplification in improving comprehension. Instruction with elaborated input should accelerate the progression to fluent reading of unmodified materials, which is the ultimate goal of foreign language reading instruction.