This paper reports a study into the role of individual (S, P, N, C, Si, Mn, Ni, Cr) and combined (Q1=S, Q2=S+P, Q3=S+P+N, Q4=S+P+N+C, Q5=S+P+N+C+Si, Q6=S+P+N+C+Si+Mn, Q7=S+P+N+C+Si+Mn+Ni, Q8=Q=S+P+N+C+Si+ Mn+Ni+Cr) elements in five steel AISI 304 smelting cycles. The correlation between the rate of corrosion K in chloride-containing media and the specific magnetic susceptibility c0 of austenite (matrix), the low content Рdof d-ferrite, and the percentage of elements has been established. Taking into consideration the order of arrangement and influence of other present components, a set of the different-shaped graphic models of K dependences on c0, Рd, and percentage of elements was found. However, the sum of the eight calculated individual and combined elements (Q8) of the models coincides with the sum of the same elements (Q=Q8) of steel smelting samples that were subjected to experimental measurements of c0 and Рd. The curves of the reported models were compared with experimental dependences of K on c0, Рd. The positive and negative role of individual and combined elements in the process of pitting resistance of steel smelting cycles has been identified. Given this, it is assumed that the effect exerted on K by individual and combined elements in the intervals before and after their critical content may be ambiguous. Hence, one value K can correspond to several values of the contents of elements, c0, Рd. A proof of that is the coincidence between the calculated models K of corrosion on the same total content of Q8 for steel samples determined experimentally. The positive (negative) and ambiguous role of elements in the process of corrosion and the possibility of predicting corrosion tolerance of austenitic steels are assumed. The experimental dependence K on c0 and Рd has been established; the greater c0 and Рd, the lower the corrosion rate K. The studied steels contained d-ferrite in the low limits of 0.01...0.1 %.