# `LlamaCppEx.Server.Strategy.PrefillPriority`
[🔗](https://github.com/nyo16/llama_cpp_ex/blob/main/lib/llama_cpp_ex/server/strategy/prefill_priority.ex#L1)

Prefill-priority batching strategy.

Prefill chunks are added to the batch first, decode tokens fill the
remaining budget. This prioritizes getting new requests through prefill
quickly, which is optimal for batch processing workloads where overall
throughput matters more than per-request generation latency.

---

*Consult [api-reference.md](api-reference.md) for complete listing*