In February, Inference billing usage had been a fixed rate while we added pay-as-you-go billing support. Starting in March, usage now takes into account compute time x price of the hardware. We’re really sorry for any confusion!
We have more information about Inference Providers here: Inference Providers.