Infant language acquisition is marked by rapid learning, with comprehension of some basic words emerging within the first year of life. Measuring the earliest stages of lexical acquisition experimentally remains challenging. In this study, we assess the acquisition of lexical knowledge from electrophysiological (EEG) recordings of infants during natural language processing. Our dataset comprised a final sample of 51 children between 3 months to 5 years of age, assessed twice within a 3-month time window. Children heard stories in their native language (German) and an unfamiliar language (French). Electrophysiological responses to the individual words in the story were quantified using EEG encoding models. We observed native-specific electrophysiological responses to words, which increase with age (t = 2.34, p = .021). To estimate whether word onset responses are related to lexical processing, we employed LLMs (GPT-3.5) to generate ageappropriate text corpora and estimate lexical frequencies and contextual lexical predictabilities of the words in our stimuli. Amplitudes between 100–300 ms increase with word frequency (t = 3.14, p < .001), an effect attenuated by age (t = –2.22, p = .026). Amplitudes between 300–450 ms decrease with word frequency (t = –2.26, p = .024). Between 450–600 ms, amplitudes are modulated by age-appropriate lexical predictability (t = –4.88, p < .001). Thus, our findings reveal early electrophysiological sensitivity to individual words and words in context, suggesting that age-appropriate LLMs may be a critical looking glass for studying the early emergence of language comprehension with non-invasive electrophysiological recordings.