"NVIDIA NIM, part of NVIDIA AI Enterprise, is a set of accelerated inference microservices that allow organizations to run AI models on NVIDIA GPUs anywhere" Nvidia
Try API's and models explorer, with some free credits even without registration: https://build.nvidia.com/explore/discover
Github examples: https://github.com/NVIDIA/GenerativeAIExamples
Nvidia NIM: https://developer.nvidia.com/nim
In my Jupyter notebook you can try NVIDIA NIM ligthining fast stabilityai/stable-diffusion-xl API for text-to-image with API key from https://build.nvidia.com/explore/discover there also try models with no keys and registrations.