Chemical recycling of plastic wastes through pyrolysis, gasification, or partial oxidation is a promising alternative to landfill disposal and incineration which must be applied in a future circular economy. These technologies enable the chemical industry, which currently heavily relies on crude oil, to obtain necessary chemical feedstock from postconsumer plastic waste. Kinetic models of pyrolysis and gasification reactions are required to dimension and design these processes on an industrial scale. The creation of detailed kinetic networks is often not feasible due to their complexity in this application, which is when the lumped kinetic modeling approach is used. This work develops and compares five lumped kinetic models for the co-pyrolysis of LDPE with a heavy petroleum fraction in a tubular reactor. A priori lumping is used for four models, and the fifth is created using a posteriori principle whereby in each model the product mixture is defined by eight lumps distinguished by their boiling point. The aim of this work is to compare different approaches for modeling reaction pathways in lumped kinetic models and to identify their impact on the predictive accuracy of the model. It was shown that all of the modeling approaches and the resulting models have similar prediction accuracies and deviations but with different kinetic parameters. Each model was used for a scale-up of an industrial-sized reactor to check whether the model had an influence on the design or predicted operation of the reactor.