This thesis aims to study the performance differences between WebGL and WebGPU using the libraries Three.js and Babylon.js through an experimental approach. By developing an e-commerce website hosting both multiple and individual 3D models, external scripts were used to measure the loading and rendering time as well as the frames per second. The frames per second were measured by removing the frame limitation over a period of 20 minutes and the loading and rendering times were measured by repeatedly reloading the website and measuring how long it took, over the course of 300 measurements. The findings from this experiment shows that Babylon.js using WebGPU does in fact not present any superiority in terms of performance, where Three.js using WebGL consistently performed better than the other combinations of API and library. Given these foundational studies showing that WebGPU performs better than WebGL in other contexts, further research is needed to understand in which contexts each API excels and if so, what optimizations and specifications are necessary for WebGPU to shine.