Abstract Despite a long-standing awareness of academic language as a pedagogically relevant research area, the construct of academic-language proficiency, understood as a more comprehensive set of skills than just academic vocabulary, has remained vaguely specified. In this study, we explore a more inclusive operationalization of an academic-language proficiency construct, core academic-language skills (CALS). CALS refers to a constellation of high-utility language skills hypothesized to support reading comprehension across school content areas. Using the Core Academic Language Skills Instrument (CALS-I), a theoretically grounded and psychometrically robust innovative tool, we first examined the variability in students' CALS by grade, English-proficiency designation, and socioeconomic status (SES). Then, we examined the contribution of CALS to reading comprehension using academic vocabulary knowledge, word reading fluency, and sociodemographic factors as covariates. A linguistically and socioeconomically diverse cross-sectional sample of 218 students (grades 4–6) participated in four assessments: the CALS-I, a standardized reading comprehension assessment (Gates–MacGinitie Reading Test), an academic vocabulary test (Vocabulary Association Test), and a word reading fluency test (Test of Silent Word Reading Fluency). General linear model analysis of variance revealed that CALS differed significantly by grade, English-proficiency designation, and SES, with students in higher grades, English-proficient students, and those from higher SES backgrounds displaying higher scores, on average. Hierarchical multiple regression analyses identified CALS as an independent predictor of reading comprehension, even after controlling for academic vocabulary knowledge, word reading fluency, and sociodemographic factors. By specifying a set of language skills associated with reading comprehension, this study advances our understanding of school-relevant language skills, making them more visible for researchers and educators.