Frankly, if I were getting a new logo made for anything serious, I’d want it in vector format. We don’t presently have any models trained to create vector art that I’m aware of, just raster images of vector art, and vectorization of raster images isn’t really good enough today to use that as a path to vector output.