The paper analyses how four screen activities relate to reading scores using the representative sample of 21,217 ten-year-olds who sat online and offline Progress in International Reading Literacy Study (PIRLS) test in six highincome European countries. In regression models, gaming and Internet use showed a right-skewed inverted Ushape relationship to online reading with moderate use (30–60 min daily) showing a positive association when compared to both no-use and heavy use (above 2 h). Online chatting and watching videos showed negative relationship to online reading above the threshold of approximately one hour daily. Quantile regression showed that all four types of screen time had similar influence on top and bottom performers except for gaming over 2 h daily which was associated with 26-point (or over a quarter of a standard deviation) lower score for lowperformers and 6-point lower score for top-performers. The paper documents the emergence of online-offline reading gaps: children who reported no screen use scored 6–11 points lower on online than offline test. Similarly, children who spent more time online scored higher on online tests than on offline tests. Whenever the heavy screen use yielded significant results, it was associated with higher online score (ranging from 8 to 16 points) when compared to offline score. A common finding for all screen activities, testing modes and groups of performers is an adverse effect on reading of more than two hours daily of screen time.