---
title: "Speculative decoding"
description: "Decode faster without changing the output: a small draft model proposes several tokens, the big model verifies them in one parallel pass, accepting the pre"
canonical_url: "https://fanout.sh/ai-engineering/lessons/speculative-decoding"
md_url: "https://fanout.sh/ai-engineering/lessons/speculative-decoding.md"
access: "public"
---

# Speculative decoding

Decode faster without changing the output: a small draft model proposes several tokens, the big model verifies them in one parallel pass, accepting the pre

## Public overview

Decode faster without changing the output: a small draft model proposes several tokens, the big model verifies them in one parallel pass, accepting the prefix that matches its own distribution. Exactness preserved, 2-3× latency wins when the draft model guesses well.

---
This representation contains public Fanout content only. Protected Pro lessons, account data, billing, checkout, and pricing are not included.

Browse the public content map: https://fanout.sh/sitemap.md