The initiator efficiency of organic peroxides in the LDPE (low-density polyethylene) process is affected by mixing. Models that more accurately consider the mixing behavior of LDPE reactors can better predict the initiator efficiency and, consequently, the polymer properties, thereby reducing the need for parameter adjustment. In this context, CFD (Computational Fluid Dynamics) simulations are an important modeling approach and form the basis of so-called compartmentalization models, i.e., networks of continuously stirred tank reactors (CSTRs). This work presents a CFD model of an LDPE autoclave reactor. When combined with a stochastic Monte Carlo simulation, the model can predict the influence of mixing on polymer properties such as branching densities and the molar mass distribution. The Monte Carlo simulation requires reaction probabilities for each elementary reaction in a location-dependent manner. To account for this based on the CFD simulation, the reactor is divided into regions. This partitioning of the reactor also enables the creation of a compartmentalization model that can be used independently of the feed conditions for the specified reactor geometry.