Many studies have shown that children with autism spectrum disorders (ASD) and attention-deficit/hyperactivity disorder (ADHD) have had early indicators of language delay. The aim of the present study was to examine the cognitive profile of school age children referred to a specialist clinic for ASD, ADHD, or both, and relate this profile specifically to the age at which these children were first flagged up (or not) as suspected from language delay during the preschool years. Forty clinic children with ASD, ADHD, or the combination of the two (without clinical suspicion of learning disability) were assessed cognitively and as regards language development and language function at a mean age of 7.3 years. They were contrasted with a group of 21 children from the community who had been flagged at 2.5 years as suspected of language delay, and who had been followed up neuropsyhiatrically/neuropsychologically and in respect of language at a mean age of 7.9 years. Mean WISC-III full scale IQ was lower than population norms (in spite of the exclusion in both samples of cases with obvious learning disability) and similar across diagnostic groups (ASD and ADHD), and across settings (clinic and community). WISC-III Kaufman factor profiles separated the diagnostic groups as regards Perceptual Organisation. Early concern about language delay was a strong predictor of lower IQ and of distinguishing between "pure" cases of ASD and ADHD. School age clinic children who present with ASD and ADHD have a similar cognitive and early language development profile as do those children from the community, followed prospectively, who present with a suspicion of early preschool language delay and are shown at school age to suffer from ASD or ADHD. Concern about early language delay in the preschool age should prompt assessments (psychiatric and cognitively) for ASD and ADHD in a multidisciplinary setting much more often than is currently the case. In many cases early language delay, even in the absence of clear learning disability should be taken as a signal that - regardless of specific diagnosis - intellectual functioning might be in the low average range.