ModelRefs / Lookahead Decoding — AI Glossary

Lookahead Decoding — AI Glossary

A parallel decoding strategy that generates and verifies token n-grams in parallel without a separate draft model. Achieved 2–4× speedup on LLaMA-2 chat.

Overview

Lookahead decoding (Fu et al. 2023) maintains a Jacobi decoding window, iteratively refining multiple token positions simultaneously. Unlike speculative decoding it needs no separate model. Achieved 2–4× speedup on LLaMA-2 chat. Useful when model size precludes finding a suitable draft model.

Reference details

Topicinference
Also known asJacobi decoding
Last reviewed2026-06-24

Commonly confused with

Like speculative decoding, this speeds up generation without changing the output distribution. The difference is what it needs: speculation requires a separate small draft model with a matching tokenizer, which you must host and keep aligned; lookahead generates and verifies candidate n-grams using the same model, so there is no second model to serve and no acceptance rate that depends on model-pair quality.

Continue your research

Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Lookahead Decoding — AI Glossary.

Frequently asked questions

What is Lookahead Decoding?

A parallel decoding strategy that generates and verifies token n-grams in parallel without a separate draft model.

Is Lookahead Decoding the same as Jacobi decoding?

Yes — Jacobi decoding are common aliases for Lookahead Decoding.

What concepts are related to Lookahead Decoding?

Closely related concepts include speculative decoding, draft model, throughput.