Which combination BEST optimizes cost for a nightly document classification pipeline?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Nightly jobs: cache the shared prompt, batch the volume, pick the smallest model that still hits quality.
Full explanation below image
Full Explanation
Batch + caching + right-sized models cut cost for offline pipelines. Always-Opus streaming is wasteful for nightly bulk.