Multi-parametric diffusion models describe non-Gaussian diffusivity of tissue observed by quantitative diffusion weighted imaging (qDWI). The ground-truth values for assumed model parameters are useful for both testing the analysis algorithms and optimizing qDWI acquisition. These values can be provided by virtual phantoms, termed digital reference objects (DROs), based on forward modeling. Here we describe a tool to generate DWI DICOM DROs with varying diffusivity model parameters and b-values including physical noise. Application examples are shown for Kurtosis, Stretched Exponential and Perfusion-fraction Intra-Voxel Incoherent-Motion DWI DROs to quantify noise-induced bias and variation in derived qDWI parametric maps.