The computing power available to run an AI model for the user. Inference is an AI term that refers to the software that generates the answers. The inference "engine" uses an AI model that has been trained on massive amounts of general data or specific data depending on the intended use. See
inference engine. Contrast with
training compute.