Large-eddy simulation (LES) can be a very important tool to support and accelerate the energy transition to green technologies and thus play a significant role in the fight against climate change. However, especially LES of reactive flows is still challenging, e.g., with respect to emission prediction, and perfect subfilter models do not yet exist. Recently, new subfilter models based on physics-informed generative adversarial networks (GANs), called physics-informed enhanced super-resolution GANs (PIESRGANs), have been developed and successfully applied to a wide range of flows, including decaying turbulence, sprays, and finite-rate-chemistry flows. This technique, based on AI super-resolution, allows for the systematic derivation of accurate subfilter models from direct numerical simulation (DNS) data, which is critical, e.g., for the development of efficient energy devices based on advanced fuels. This paper describes a case study demonstrating PIESRGANA for a finite-rate chemical methane jet flow using transfer learning. A priori and a posteriori results are presented and discussed. Since the training process is very crucial for the successful application of this new LES technique, a detailed description of possible strategies is provided.