The reaction between C₂ cluster and C₆₀ fullerene resulting in C₂ insertion to C₆₀ with formation of closed C₆₂ cage (reaction of C₂ ingestion by C₆₀) was investigated by the semiempirical MNDO-PM3 method. The geometries and energies of extremal points on the C₆₂ potential energy surface were calculated. Several reaction pathways leading to the formation of three different closed C₆₂ fullerenes were investigated. All insertion reactions proceed stepwise through intermediate adducts of different structures. The main reaction pathways were found to be addition of C₂ by its one side to the 6,6- or 5,6-bond of C₆₀ with formation of primary unclosed C₆₂ adducts of "ball-with-fork" structures, lying in deep potential wells. Back reaction of C₂ detachment from primary adducts can compete with that of their transformation to the closed C₆₂ cages inasmuch as calculated activation barriers of the both reactions are comparable. Model calculations at the B3LYP/6-31G level, using C₃₂H₁₂ semisphere instead of C₆₀, confirmed the conclusion about two competitive pathways of the primary adducts transformation, C₂ detachment, and C₂ ingestion. The concerted insertion of C₂ to C₆₀ was realized only in the case of severe restrictions on starting geometry of the C₂ + C₆₀ system. The results of calculations explain recent experimental data on the formation of metastable adducts upon addition of C₂ to C₆₀, obtained using the time-of-flight mass spectrometer with laser desorption.