Throughout the various stages of aircraft design, an accurate evaluation of performance and flight characteristics is vital, demanding extensive and precise aerodynamic data. Recent advancements in computational fluid dynamics (CFD) combined with the increased availability of high-performance computing (HPC) resources have facilitated the adoption of high-fidelity CFD solutions even in conceptual and preliminary design phases. In the case of combat aircraft that operates in diverse flight regimes involving high angles of attack and leading-edge vortices, the use of highly accurate flow solutions becomes imperative. However, the significant manual effort required for creating suitable mesh discretisation of geometries in early design stages and the computational cost associated with obtaining such datasets pose limitations. Nevertheless, for aerodynamically less demanding flight phases, simpler and faster aerodynamic simulation tools suffice for conceptual design studies. Integrating disparate data sources with differing accuracies and uncertainties, depending on the specific flight conditions of interest, can be challenging. Multifidelity modelling offers a route to enhance the efficiency and reduce the cost of aerodynamic analysis and design processes by incorporating data from multiple sources. There is a growing trend towards machine learning-based techniques for building multifidelity surrogate models, including Gaussian processes, neural networks, and dimensionality reduction methods. Machine learning tools offer practical and efficient means of combining information from multiple data sets, allowing for the integration of real-world variability and probabilistic behaviour into engineering analysis and design processes. In this work, two multifidelity approaches based on Gaussian processes are applied and compared. The popularity of Gaussian process surrogate modelling stems from its ability to characterise multidimensional surfaces, while accounting for the correlation of the data in the design space and providing a measure of confidence in the predictions. Both methods are implemented in the SMARTy framework, a Python API developed by DLR for data-driven applications in the aerospace sector. The objectives of this collaboration between the University of Liverpool and the DLR is to compare these approaches on a challenging test case and to assess their respective accuracy and output uncertainty.