TOPIC #171Advanced 14 min read

Planning: Chain-of-Thought, Tree-of-Thoughts & Graph-of-Thoughts

AI
AI & ML Editorial
Report an issue
Key takeawayCore Concept Summary

How LLMs deliberate: Chain-of-Thought writes one reasoning line, Tree-of-Thoughts explores branches and backtracks, Graph-of-Thoughts merges many partial answers. This topic covers how each works, what each costs, and what reasoning models changed in 2024–2025.

01.The Problem: The Model's First Thought Can Be Wrong

Give a model a puzzle:

Insight

"Make 24 using the cards 4, 9, 10, 13 — you may use +, −, ×, ÷ and parentheses."

Ask it directly and it blurts one answer in a single pass. If its first internal step is wrong, there is no way to take that step back. It can't even see that it was wrong — it just keeps generating, rationalizing its own mistake.

That is the structural weakness of plain generation:

  • One shot. The answer is decoded greedily from the question.
  • No undo. Wrong early steps get rationalized, not reverted.
  • No alternatives. The model never asks "what if I tried a different approach first?"

So the question becomes

Insight

Can we make the model deliberate — write out steps, consider alternatives, check itself — before committing?

The last three years of prompting research is one long answer to that question, organized by the shape of the deliberation:

  • a line (Chain-of-Thought),
  • a tree you search (Tree-of-Thoughts),
  • a graph you merge (Graph-of-Thoughts).

This topic walks up that ladder — and ends with the 2024–2025 twist: reasoning models moved all of this inside the model.

Reasoning Topologies

PRO Architecture Blueprint

Reasoning Topologies

CoT samples one path; ToT searches a tree with BFS/DFS and value functions; GoT merges and refines parallel thoughts like a DAG operation.

Reasoning Topologies
100%
Touchpad: Pinch to zoom • Drag to pan
Rendering visual architecture flowchart...
PRO & LIFETIME CURRICULUM

Unlock Topic #171: Planning: Chain-of-Thought, Tree-of-Thoughts & Graph-of-Thoughts

You are viewing a preview. The full in-depth technical walkthrough, worked derivations, and code notebooks for this concept, along with self-assessment quizzes, are available with Pro or Lifetime Access.

Production Deep Dive

Failure modes, high-throughput bottlenecks, and real FAANG implementation decisions.

Interactive Blueprints

Interactive system topology diagrams, live parameter simulators, and downloadable SVG charts.

Knowledge Assessment

Staff-level multiple-choice quiz questions with instant feedback and answer explanations.

Cross-Device Progress Sync

Firebase Google authentication automatically syncs your completed topics and quiz scores.

Rate This Architecture ChapterFeedback & Rating

How clear and actionable was this distributed systems breakdown?