That type 1 diabetes (T1D) arises as a result of the concerted efforts of the adaptive immune system to destroy β-cells is axiomatic, yet the relationship between autoreactive T cell activation, autoantibodies and β-cell loss is ill-defined. Recent data suggest that T1D is a heterogeneous disorder both in terms of disease progression and in the lack of uniformity in the underlying pathology of the disease. We set out to examine whether the adaptive autoimmune response can provide insight into such heterogeneity. The frequencies of peripheral blood CD4 T cell responses (IFN-γ and IL-10) to multiple β-cell autoantigenic peptides and autoantibodies (AAbs) were studied in 33 children with new-onset T1D and at least one DRB1*0301 or 0401 allele. A further 72 children (non-diabetic siblings of T1D patients and possessing DRB1*0301 or 0401) were studied. In the T1D patients we used the combined dataset (27 analytes) in unbiased, hierarchical cluster analysis, revealing two equal-sized, distinct agglomerations (bootstrap support for the main nodes 88-97%). One is characterised by IFN-γ reactivity and multi-AAb-positivity (“inflammatory autoimmune phenotype”) and the other by IL-10 reactivity (“partially-regulated autoimmune phenotype”). Compared with siblings, IFN-γ reactivity to β-cell autoantigens, and especially insulin/proinsulin, was strongly and significantly associated with T1D. We then imposed the T1D clustering model in 15 high risk unaffected siblings with ≥2 AAbs positive and observed the same agglomerations, suggesting that the inflammatory and partially regulated autoimmune phenotypes may precede T1D. These studies support the idea of heterogeneity in T1D and argue that at least two distinct immunopathological processes may underlie disease development, providing an opportunity for stratification in patient selection for prevention and intervention studies.