Top-N AI model recommendation for a task. Joins live pricing, benchmarks, provider status, and latency into a single composite score so an agent gets a ranked answer in one call instead of stitching three free endpoints together.