Existing studies on SARS-CoV-2 infection have mainly focused on severe clinical outcomes; understanding of the clinical severity profile of general community infections is poor. We aimed to assess and compare the clinical profiles of infections with SARS-CoV-2 omicron (B.1.1.529) subvariants in a representative community cohort in Hong Kong during periods of BA.2, BA.4/5, and XBB dominance. In this prospective cohort study in Hong Kong, a representative community cohort of individuals aged at least 5 years were recruited by random-digit dialling and underwent weekly rapid antigen testing for SARS-CoV-2, irrespective of symptoms, during three periods from March 1, 2022, to Oct 31, 2023, in which the BA.2, BA.4/5, or XBB subvariants were dominant. We analysed the likelihood of symptoms, as well as the patterns, severity, and duration of symptoms and their associations with participant demographics and vaccination and infection histories. 1126 (11·0%) of 10 279 participants in the BA.2 period, 830 (6·6%) of 12 588 in the BA.4/5 period, and 633 (11·1%) of 5690 during the XBB period tested positive for SARS-CoV-2 infection on rapid antigen tests. Community infections were generally mild, with asymptomatic infections comprising 22·0-25·0% of infections. No hospitalisations or deaths occurred as a direct result of SARS-CoV-2 infection during the study period. Compared with children aged 5-17 years, a higher likelihood of being symptomatic on infection was found for adults aged 18-59 years during the period of BA.2 dominance and adults aged 60 years or older during XBB dominance. Most (>90·0%) participants with symptomatic infections reported respiratory and systemic symptoms. Up-to-date vaccination with a regimen containing the BNT162b2 vaccine, compared with those without an up-to-date vaccine, was associated with a reduced likelihood of symptoms on infection during the period of BA.2 dominance and of severe symptoms causing substantial disturbance to daily life (grade 3 symptoms) during periods of BA.2 and BA.4/5 dominance, whereas no association was observed during the period of XBB dominance. Previous SARS-CoV-2 infection was associated with a reduced likelihood of symptoms on infection during BA.4/5 and XBB dominance and of severe symptoms during XBB dominance. Reports of severe symptoms increased over the three periods, from 236 (27·7%) of 852 symptomatic participants during BA.2 dominance to 176 (37·1%) of 475 during XBB dominance. The duration of symptoms was longest in the BA.2 period (median 10·0 days [95% CI 9·0-10·0]) and similar in the other two periods (8·0 [8·0-9·0] during BA.4/5 dominance and 8·0 [8·0-9·0] during XBB dominance). A symptom duration of 60 days or longer was reported only during the period of BA.2 dominance, in six (0·7%) of 824 infections. SARS-CoV-2 infections were generally mild, but not increasingly so, along the evolution of omicron subvariants in this highly vaccinated population. About a third of participants with symptomatic infections reported that the symptoms severely affected daily life even if they were not admitted to hospital, resulting in morbidity, absence from work or school due to illness, productivity loss, and increased medicoeconomic burden. A gradual reduction in the association of vaccines and increase in the association of previous infection with the symptom profile, possibly reflecting the effects of immune escape and waning, were observed over the study period. Henry Fok Foundation and Hong Kong Health Bureau.