Last updated: 2026-04-18
What is Batching?
Processing N items in one LLM call instead of N separate calls. Halves per-item overhead and pairs perfectly with prompt caching. Typical batches: 10 text items or 5 long transcripts per call.
See also
← Back to the full AI agent glossary.