STUDY QUESTION: When do the differences in birth weights become apparent between singletons born after frozen embryo transfer (FET) and fresh embryo transfer (fresh ET)? SUMMARY ANSWER: Mean birth weights after FET become significantly higher starting from gestational week (GW) 33 among boys and from GW 34 among girls. WHAT IS KNOWN ALREADY: In recent years, there has been a steep rise in recorded FET treatments, enabling widespread use of elective single embryo transfer, thus reducing the risks associated with multiple gestations. However, singletons born after FET are heavier and there is a higher risk of large-for-gestational-age (LGA) (birth weight > 90 percentiles) compared to fresh ET. In contrast, risk of small-for-gestational-age (SGA, birth weight < 10 percentiles) is lower in singletons born after FET compared to fresh ET. The reasons, timing and consequences of these differences remain largely unclear. There is limited evidence about whether this difference in growth develops before the last trimester of pregnancy. STUDY DESIGN, SIZE, DURATION: This retrospective Nordic register-based cohort study compared singletons born after FET (n = 17 500) to singletons born after fresh ET (n = 69 510) and natural conception (NC, n = 3 311 588). All live born singletons born between the years 2000 and 2015 in Denmark, Norway and Sweden at gestational age ≥22 weeks were included from the population-based Committee of Nordic ART and Safety (CoNARTaS) study population. PARTICIPANTS/MATERIALS, SETTING, METHODS: Children born after FET were compared to those born after fresh ET and NC for mean birth weight and proportion of LGA and SGA for each GW at birth. Chi-square test and tests for relative proportions were used to compare categorical variables and Student's t-test was used to compare continuous variables. Adjusted odds ratios (aORs) for LGA and SGA were calculated using logistic regressions, adjusting for year of birth, maternal age, parity, BMI, chronic hypertension, diabetes, smoking and offspring sex. MAIN RESULTS AND THE ROLE OF CHANCE: Mean birth weights were significantly higher after FET compared to fresh ET starting from GW 33 (range from 75 g to 228 g by week) for boys and starting from GW 34 (range from 90 g to 236 g by week) for girls. Boys born after FET had a significantly higher proportion of LGA (11.0-15.1%) at birth between GW 36 and 42, compared to those born after fresh ET (7.1-9.4%) (range from P < 0.001 to P = 0.048 by week). For girls born after FET, the difference was seen between GW 37 and 42 (10.6-13.4%) compared to those born after fresh ET (6.6-8.0%) (range from P < 0.001 to P = 0.009 by week). The proportion of SGA was significantly lower among boys born after FET (7.6-8.7%) compared to fresh ET (11.9-13.6%) between GW 36 and 42 (range from P < 0.001 to P = 0.016 by week). For girls born after FET, the difference was seen between GW 38 and 42 (7.0-9.3%) compared to those born after fresh ET (13.0-14.6%) (P < 0.001). The proportion of LGA (12.3-15.1%) was significantly higher for boys born after FET between GW 38 and 41 (P < 0.001) and for girls born after FET (12.6-13.4%) between GW 37 and 40 (range from P < 0.001 to P = 0.018 by week), compared to naturally conceived boys (9.7-9.9%) and girls (9.0-10.0%). All singletons born after FET had a higher risk of LGA compared to singletons born after fresh ET (aOR 1.87, 95% CI 1.76-1.98) and singletons born after NC (aOR 1.28, 95% CI 1.22-1.35). LIMITATIONS, REASONS FOR CAUTION: There may be residual confounding factors that we were not able to control for, most importantly the causes of preterm birth, which may also influence foetal growth. A further limitation is that we have no knowledge on growth patterns between implantation and GW 22. Finally, the number of children born extremely preterm or post-term was limited even in this large study population. WIDER IMPLICATIONS OF THE FINDINGS: This is, to date, the largest study on birth weights among preterm and term ART singletons with a population-based design and NC control group. The results suggest that the freeze-thaw process is associated with higher birthweights and greater risk of LGA at least in the last trimester of pregnancy. This is an important aspect of the safety profile of ART. More research is needed on the long-term outcome of these children. STUDY FUNDING/COMPETING INTEREST(S): The CoNARTaS collaboration has received the following funding: the Nordic Trial Alliance: a pilot project jointly funded by the Nordic Council of Ministers and NordForsk [71450], the Central Norway Regional Health Authorities [46045000], the Norwegian Cancer Society [182356-2016], the Nordic Federation of Obstetrics and Gynaecology [NF13041, NF15058, NF16026 and NF17043], the Interreg Öresund-Kattegat-Skagerrak European Regional Development Fund (ReproUnion project) and the Research Council of Norway's Centre of Excellence funding scheme [262700]. None of the authors have any competing interests to declare. TRIAL REGISTRATION NUMBER: ISRCTN11780826.