Inference is a global network of compute providers delivering affordable, serverless inference for the top open source AI models. We built a distributed infrastructure that allows developers to access state-of-the-art language models with the reliability of major cloud providers but at a fraction of the cost.