Recent efforts have defined developmental trajectories for diverse neuroimaging biomarkers, allowing for the detection of deviation from normality on an individual level. The functional connectome captures how brain areas communicate and graph theory analyses can characterise this at a network level. We modeled typical developmental trajectories of the functional connectome using graph theory to explore network-level development during childhood and adolescence. Using resting state functional MRI, we measured functional connectivity between 106 ROIs (Harvard-Oxford atlas), in a cross-sectional sample of typically developing participants (N=219, aged 6-24 years). Functional connectomes were quantified using a weighted, directed graph analysis and connection density, modularity, clustering coefficient, global efficiency, and betweenness centrality were calculated separately for positive and negative connectivities. For positive connectivities, modularity and betweenness centrality increased with age (both p<0.001), while connection density, clustering coefficient, and global efficiency decreased with age (all p<0.001). By contrast, for negative connectivities, modularity, and betweenness centrality decreased with age (p=0.002, p=0.003), while connection density, clustering coefficient, and global efficiency increased with age (p<0.001, p<0.001, p=0.003). See figure below. Developmental changes may be driven by an intrinsic pressure to balance functionality with low metabolic cost to maintain the network. The positive connection network appears to shift towards a more efficient conformation resembling “small-world” architecture. In contrast, the negative connection network seems to shift away from such efficient architecture, possibly to prioritize improving functionality before refinement in later adolescence or early adulthood.