Simulations that combine flow modeling with radionuclide transport modeling provide insights into the performance of a hypothetical geologic repository for radioactive waste. Key parameters driving performance can be identified and quantified using probabilistic sampling of the model parameters. This paper presents a probabilistic framework, referred to as the MC-MELODIE hydrologic flow and contaminant transport model, for analyzing flow and transport behavior in the context of intermediate level long-lived (ILLLW) and high-level (HLW) disposal in a deep geological formation at the Meuse/Haute-Marne (MHM) site in France. The flow and transport framework includes nested, but separate, simulations with common parameters in order to consider small-scale features (detailed drift and shaft configurations) within the repository formation while also considering large-scale (regional) release. Specific analyses use one-million-year simulations with a conservative tracer and assuming instantaneous failures and steady flow to examine the effect of repository layout on radionuclide transfer to the surrounding aquifers through the host formation and engineered features, and to illustrate how uncertainties in far-field conditions and potential future human activities may affect the fate of released radionuclides.