The NCN radical plays a key role for modeling prompt-NO formation in hydrocarbon flames. Recently, in a combined shock tube and flame modeling study, the so far neglected reaction NCN + H2 and the related chemistry of the main product HNCN turned out to be significant for NO modeling under fuel-rich conditions. In this study, the reaction has been thoroughly revisited by detailed quantum chemical rate constant calculations both for the singlet 1NCN and triplet 3NCN pathways. Optimized geometries and vibrational frequencies of reactants, products, and transition states were calculated on B3LYP/aug-cc-pVQZ level with single-point energy calculations carried out against the optimized structures using CASPT2/aug-cc-pVQZ. The determined rate constants for the 1NCN + H2 reaction as well as the newly measured high temperature absorption cross section of 3NCN made a reevaluation of the shock tube data of the previous work necessary, finally revealing quantitative agreement between experiment and theory. Moreover, the new directly measured Doppler-limited absorption cross section data, σ­(3NCN, λ = 329.1302 nm) = 2.63 × 109 × exp­(−1.96 × 10–3 × T/K) cm2/mol (±23%, p = 0 bar, T = 870–1700 K), are in agreement with previously reported values based on detailed spectroscopic simulations. Hence, a long-standing debate about a reliable high temperature 3NCN absorption cross section has been resolved. Whereas 3NCN + H2 resembles a simple abstraction type reaction with the exclusive products HNCN + H, the singlet radical reaction is initiated by the insertion into the H–H bond. Up to pressures of 100 bar, the main products of the subsequent decomposition of the H2NCN intermediate are HNCN + H as well, with minor contributions of CN + NH2 toward higher temperatures. Although much faster than the triplet reaction, the singlet radical insertion is actually rather slow, due to the necessary reorganization of the HOMO electron density in 1NCN that is equally distributed over the two N atom sites. In general, the distinct reactivity differences call for a separate treatment of 1NCN and 3NCN chemistry. However, as the main reaction products in case of the H2 reaction are the same and as the population of the 1NCN in thermal equilibrium remains low, a properly weighted effective rate constant k(NCN + H2 → HNCN + H) = 2.62 × 104 × (T/K)2.78 × exp­(−97.6 kJ/mol/RT) cm3 mol–1s–1(±30%, 800 K < T < 3000 K, p < 100 bar) is recommended for inclusion into flame models that, as yet, do not explicitly account for 1NCN chemistry.