Finds the cheapest H200 GPU rental hourly rate across cloud providers for large model inference.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.