Building a Thinking Model from a Non-Thinking Model Using Chain-of-Thought (CoT)


Ever wondered how to make an AI not just answer but think like a human?
That’s exactly what Chain-of-Thought (CoT) prompting does.
In this article, we’ll break it down for freshers and beginners — with simple examples, visuals, and deep thoughts.
🌱 1. What is a Non-Thinking Model?
A non-thinking model is like a student who answers without showing their work.
For example:
Prompt:
What’s 27 × 14?
Non-thinking Model Output:
378
✅ Correct, but we don’t know how it got there.
No reasoning, no steps — just an answer.
💡 2. What is a Thinking Model?
A thinking model doesn’t just give the answer — it shows the reasoning steps.
It’s like a teacher explaining the process instead of only saying the result.
Prompt with Chain-of-Thought:
Let’s think step-by-step. What’s 27 × 14?
Thinking Model Output:
Step 1: 27 × 10 = 270
Step 2: 27 × 4 = 108
Step 3: 270 + 108 = 378
Final Answer: 378
Now we know exactly how the model got the answer.
🪄 3. What is Chain-of-Thought (CoT)?
Chain-of-Thought prompting is a way to make AI explain its reasoning step-by-step.
It transforms a non-thinking model into a reasoning model by guiding it to “think” before answering.
📌 Key idea: Instead of asking for the answer directly, you ask it to explain the steps first.
🧩 4. Example: From Non-Thinking → Thinking
Non-Thinking Prompt:
Who is taller, Tom (6'2") or Jerry (5'9")?
Answer: Tom.
Thinking Prompt (CoT):
Let’s think step-by-step. Compare Tom’s height (6'2") and Jerry’s height (5'9").
Answer: Tom is 6'2", Jerry is 5'9". 6'2" is greater than 5'9". So, Tom is taller.
🔍 5. Why CoT is Important
Transparency – We can verify reasoning.
Debugging – Spot where the model went wrong.
Teaching – Great for learning and tutoring applications.
Complex Problem Solving – Math, coding, logic puzzles, and planning benefit from reasoning.
🛠️ 6. How to Use Chain-of-Thought as a Beginner
You don’t need to change the model — just change your prompt.
Some starter phrases:
“Let’s think step-by-step.”
“Explain your reasoning before giving the final answer.”
“Show the calculation process.”
🔮 7. Deep Thought
AI without reasoning is like a calculator.
AI with reasoning is like a teacher.
The difference isn’t in knowledge, but in how they show the journey from problem to answer.
Subscribe to my newsletter
Read articles from Ankit Chaudhary directly inside your inbox. Subscribe to the newsletter, and don't miss out.
Written by
