Autism sibling recurrence in prospective infant family history studies is ~20% at 3 years but systematic follow-up to mid-childhood is rare. In population and clinical cohorts autism is not recognized in some children until school-age or later. One hundred and fifty-nine infants with an older sibling with autism underwent research diagnostic assessments at 3 years and mid-childhood (6 to 12 years (mean 9)). We report the autism sibling recurrence rate in mid-childhood and compare developmental and behavioral profiles at mid-childhood and 3 years in those with earlier versus later recognized autism, and those who had, or had not, received a community autism diagnosis. The autism recurrence rate in this sample in mid-childhood was 37.1%, 95% CI [29.9%, 44.9%] and higher in boys than girls. Around half of those diagnosed with autism in mid-childhood had not received a diagnosis at 3 years. Later, diagnosis was more common in girls than boys. While some had sub-threshold symptoms at 3, in others late diagnosis followed a largely typical early presentation. Sibling recurrence based on community clinical diagnosis was 24.5%, 95% CI [18.4%, 31.9%]. Those who also had a community diagnosis tended to be older, have lower adaptive function and higher autism and inattention symptoms. Notwithstanding limitations of a single site study, modest sample size and limits to generalisability, autism sibling recurrence in family history infants may be higher in mid-childhood than in studies reporting diagnostic outcome at 3 years. Findings have implications for families and clinical services, and for prospective family history studies.