ModelRefs / Beam Search — AI Glossary
Beam Search — AI Glossary
A decoding strategy maintaining B candidate sequences (beams) simultaneously, selecting the overall highest-probability completion. Also called beam decoding.
Overview
Beam search expands B candidates at each step, keeping the B highest-scoring partial sequences. Unlike greedy decoding it avoids locally optimal but globally poor choices. Historically dominant for machine translation; generally replaced by sampling (temperature + top-p) for open-ended generation because sampled outputs are more diverse and human-preferred.
Reference details
| Topic | inference |
|---|---|
| Also known as | beam decoding |
| Last reviewed | 2026-06-24 |
Related terms
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Beam Search — AI Glossary.
Frequently asked questions
What is Beam Search?
A decoding strategy maintaining B candidate sequences (beams) simultaneously, selecting the overall highest-probability completion.
Is Beam Search the same as beam decoding?
Yes — beam decoding are common aliases for Beam Search.
What concepts are related to Beam Search?
Closely related concepts include greedy decoding, sampling, temperature.