Managing risks associated with commodities is crucial to ensure that business operations leadto favorable financial results and reduce the risk of short-term financeability problems. Toachieve this, it is necessary to create scenarios for commodity prices that accurately reflect theirprobability distributions. This paper presents an implementation of the well-establishedTimeGAN architecture for generating multiple scenarios of gasoline crack spread, with theobjective of supporting risk management and business decisions. This approach offers acomplementary approach to traditional stochastic models based on Stochastic DifferentialEquations for time series simulation and risk analysis. It leverages the powerful capabilities ofGenerative Adversarial Networks (GANs) to produce realistic scenarios, particularly incapturing complex probabilistic distributions without needing any assumptions about the datadistribution. By accurately modeling the probabilistic distribution of critical risk factors, theGAN framework enables more reliable estimation of their potential impact on businessperformance, making it a robust tool for financial risk assessment.