This paper empirically investigates the dynamic behavior of non-ferrous metal prices, specifically for copper, aluminum, nickel, and zinc, using a range of continuous-time diffusion models. The study employs maximum likelihood estimation with approximate transition probability density functions to analyze daily price data. Our findings reveal that while price volatility is the dominant factor explaining the dynamics of all four metals, the best-fitting model varies by metal. The GD-GV model, with its generalized volatility function, is found to be the most appropriate for aluminum, whereas the more parsimonious CKLS model provides the best fit for copper, nickel, and zinc. This analysis demonstrates that price dynamics within the non-ferrous metal group are not homogeneous, underscoring the necessity of a flexible, metal-specific modeling approach for accurate price characterization.