Compares embedding models on retrieval benchmarks, including latency, cost, dimensions, and accuracy.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.