LLM APIs

Gateways, rented GPUs and managed model endpoints price, bill and limit work in different ways, so compare them in a fixed order instead of on a single headline number.