Previous studies have identified cytokines associated with respiratory virus infection illness outcome. However, few studies have included comprehensive cytokine panels, longitudinal analyses, and/or simultaneous assessment across the severity spectrum. This, coupled with subjective definitions of cytokine storm syndrome (CSS), have contributed to inconsistent findings of cytokine signatures, particularly with COVID severity. Here, we measured 38 plasma cytokines and compared profiles in healthy, SARS-CoV-2 infected, and multisystem inflammatory syndrome in children (MIS-C) patients (n = 169). Infected patients spanned the severity spectrum and were classified as Asymptomatic, Mild, Moderate or Severe. Our results showed acute cytokine profiles and longitudinal dynamics of IL1Ra, IL10, MIP1b, and IP10 can differentiate COVID severity groups. Only 4% of acutely infected patients exhibited hypercytokinemia. Of these subjects, 3 were Mild, 3 Moderate, and 1 Severe, highlighting the lack of association between CSS and COVID severity. Additionally, we identified IL1Ra and TNFa as potential biomarkers for patients at high risk for long COVID. Lastly, we compare hypercytokinemia profiles across COVID and influenza patients and show distinct elevated cytokine signatures, wherein influenza induces the most elevated cytokine profile. Together, these results identify key analytes that, if obtained at early time points, can predict COVID illness outcome and/or risk of complications, and provide novel insight for improving the conceptual framework of hypercytokinemia, wherein CSS is a subgroup that requires concomitant severe clinical manifestations, and including a list of cytokines that can distinguish between subtypes of hypercytokinemia.