Phonological similarity effects are biases to judge words as phonologically similar (i.e., rhyming), even if they are not. First found in rime awareness tasks in preliterates, these biases have recently also been found in proficient adult readers. In this study, we evaluated underlying phonological processing in rime judgment longitudinally, across literacy development. To this end, we created a new rime judgment task (rime; i.e., /t∙aɪ̯∙l/ - /z∙aɪ̯∙l/) with two distractor conditions that varied in size of phonological overlap (body; i.e., /t∙aɪ̯∙l/ - /t∙aɪ̯ ç/; nucleus; i.e., /t∙aɪ̯∙l/ - /r∙aɪ̯∙s/). The task was administered to a group of 61 German-speaking children at four time points across school entry and to 21 adults. Accuracy and latency responses were recorded. Results indicated that children and adults showed phonological similarity effects but the effect decreased gradually over time. However, preliterate children were more sensitive to large compared to small phonological overlap, while the same effect was significantly smaller in literate children and adults. Results suggest that preliterate children are more sensitive to larger grain sizes and become more sensitive to fine-grained units across literacy development. The findings are in line with the assumptions of the psycholinguistic grain size theory.