20 ChatGPT Prompts for Learning Maths Effectively

A model is a genuinely good maths tutor for one specific thing: explaining why something works. It is an unreliable calculator. Those two facts together decide how you should use it, and they are the reason this page starts with a warning rather than a prompt.

The twenty prompts below are arranged in five stages, from meeting a concept for the first time through to exam preparation and mathematical thinking. Each has a note explaining why it works, because the reasoning is more transferable than the prompt itself.

They work in ChatGPT and any comparable model. For studying more broadly, see our ChatGPT prompts for students, and for prompting technique across other work, the main ChatGPT prompts guide.

• Check every number yourself. Models make arithmetic and algebra slips, and they present them with exactly the same confidence as correct working. Trust the explanation of the method; verify the calculation independently. A wrong worked solution learned confidently is worse than no help at all.
• Do not submit anything it produced. Using it to understand a method is learning. Handing in its solution breaks almost every school and university policy, and it skips the part where you actually get better at maths. Check what your institution's rules say before you use any of this on assessed work.
• Attempt it first, always. The struggle is where the learning happens. When you do ask for help, ask for the minimum that unblocks you rather than the full solution. Prompt 7 exists for exactly this.
• Be explicit about notation. Fractions, exponents and integrals are ambiguous in plain text. State how you are writing them, or a misread symbol will produce a confidently wrong answer to a different question.
• Get it to quiz you. After any explanation, ask it to test your understanding. Reading an explanation feels like learning and often is not, and a single question will tell you which one just happened.

These prompts are for when you encounter a concept you do not understand or want to understand more deeply.

Prompt 1: Explain a Concept from Scratch

Explain [concept: e.g. fractions, quadratic equations, derivatives, probability] to me as if I have never encountered it before. Start with why this concept exists and what problem it solves. Then build up the explanation from the simplest possible idea, using a real-world analogy before introducing any notation or formula. Check my understanding at the end with one question.

Why it works: starting with why a concept exists before showing how it works is the single most effective change you can make to a maths explanation. Most textbooks do this backwards.

Prompt 2: Multiple Explanations

I have read an explanation of [concept] but it has not clicked yet. Give me three completely different ways to explain this concept: (1) using a visual or spatial analogy, (2) using a real-world practical example, and (3) using the formal mathematical definition broken down into plain language. I want to find the explanation that clicks for me.

Why it works: different people have different conceptual entry points. Having three explanations means you are much more likely to find the one that makes the concept snap into place, and exposure to all three deepens understanding regardless.

Prompt 3: Connect to What You Already Know

I understand [concept A: e.g. multiplication, linear equations, basic probability] reasonably well. Explain [concept B: e.g. exponents, quadratic equations, conditional probability] by building directly on what I know about [concept A]. Show me the connection between them and explain exactly where the new complexity is introduced and why.

Why it works: mathematical concepts build on each other in specific ways. Making those connections explicit closes the gaps that cause confusion later. This is how expert maths teachers think but most textbooks do not show it.

Prompt 4: Why Does This Rule Work?

I have been taught the rule that [state a rule: e.g. when you divide fractions you flip the second one and multiply, a negative times a negative equals a positive, the derivative of xⁿ is nxⁿ⁻¹]. I can apply this rule but I do not understand why it is true. Prove to me that this rule is correct using a simple numerical example, and then explain the underlying reason so I genuinely understand it rather than just memorising it.

Why it works: rules without reasons are fragile. When you understand why a rule is true, you can reconstruct it if you forget it, apply it more flexibly, and spot when it does not apply.

Prompt 4 is worth using on every rule you were told to memorise. If a model cannot show you why a rule is true with a small numerical example, that is a signal to check the rule itself rather than to accept the explanation.

These prompts are for when you are actively solving problems and want to learn from the process rather than just get the answer. Prompts 6 and 7 are the two that protect the learning, and they are the ones to default to.

Prompt 5: Step-by-Step with Explanations

Solve this problem step by step: [paste problem]. For each step, do not just show the calculation - explain what you are doing and why you are doing it at this stage. Treat each step as a teaching moment so I can follow the reasoning, not just the arithmetic.

Why it works: this is the difference between copying a worked solution and actually learning from one. The explanations turn a sequence of calculations into a narrative you can internalise. Check the arithmetic in every step yourself, because this is the prompt where a slip is easiest to miss.

Prompt 6: Check My Working

Here is my attempt at solving this problem: [paste problem and your working]. Check my solution. If I have made a mistake, do not just tell me the correct answer - identify exactly where my reasoning went wrong, explain why it is wrong, and give me a hint that points me toward the correct approach without doing the problem for me. If my working is correct, tell me and explain whether there is a more elegant method.

Why it works: getting feedback on your own attempt is far more valuable than reading a worked solution. This prompt mimics what a good maths teacher does - it identifies your specific misconception rather than showing you the general method again.

Prompt 7: Hint-Only Mode

I am stuck on this problem: [paste problem]. Do not solve it for me. Give me only the first hint I need to make progress - just enough to get me unstuck. After I respond with my next step, give me the next hint if I need it. We will work through this problem together with you guiding me rather than solving it.

Why it works: the Socratic method is the most powerful way to learn problem-solving. Being guided to the solution yourself produces far deeper retention than reading a worked answer. It also sidesteps the accuracy problem entirely, because you are doing the calculation.

Prompt 8: Identify the Problem Type

Here is a maths problem: [paste problem]. Before we solve it, help me understand what type of problem this is. What topic or technique does it require? What are the clues in the problem that tell you which method to use? Explain the pattern-recognition process a confident mathematician would use when they first read this problem.

Why it works: one of the hardest skills in maths is knowing which technique to use. This prompt explicitly teaches that pattern-recognition skill rather than assuming you will pick it up by osmosis.

Reach for prompt 7 before prompt 5. A hint costs you five more minutes and buys you the ability to do the next problem unaided, which is the entire point.

These prompts help you turn errors into learning opportunities and address the root causes of recurring mistakes. This is the stage most self-taught learners skip and the one that produces the biggest improvement.

Prompt 9: Diagnose My Misconception

I keep making mistakes with [topic: e.g. negative numbers, simultaneous equations, integration by parts]. Here are two or three examples of problems I have got wrong: [paste examples with my incorrect working]. Diagnose what my underlying misconception is - not just what I did wrong in each question, but what fundamental misunderstanding is causing the repeated errors. Then explain how to correct it.

Why it works: recurring mistakes almost always have a single root cause. Treating symptoms by redoing the same problems is less effective than identifying and correcting the underlying misconception.

Prompt 10: Common Mistakes for This Topic

I am about to start learning [topic: e.g. trigonometry, logarithms, vectors, statistical hypothesis testing]. Before I start, tell me: what are the five most common mistakes students make when learning this topic, what misconceptions cause them, and what I should watch out for to avoid them? I want to be aware of the pitfalls before I encounter them.

Why it works: pre-emptive awareness of common errors is far more efficient than discovering them yourself. This is the kind of insider knowledge a good teacher gives you at the start of a topic.

Prompt 11: When Does This Method Fail?

I have learned [method: e.g. the quadratic formula, the chain rule, Pythagoras' theorem, BIDMAS]. I know how to apply it but I am not sure when it does not work or when I should use a different approach. Explain the conditions under which this method is valid, give me examples of cases where it breaks down or gives wrong results, and explain what to use instead in those cases.

Why it works: most maths errors come from applying a valid method in a situation where it does not apply. Understanding the scope and limits of a technique is as important as understanding the technique itself.

Keep a note of what prompt 9 tells you. A misconception diagnosed once tends to reappear months later in a different topic, and recognising it the second time is much faster than diagnosing it again.

These prompts help you consolidate learning, build fluency and prepare for assessed work. Generated practice problems are safe to use because you are doing the solving; generated answers to them are not, so attempt everything before asking for marking.

Prompt 12: Generate Practice Problems

Generate 5 practice problems on [topic] at [level: beginner / intermediate / challenging]. Start with the easiest and increase difficulty progressively. Do not show the answers yet - I will attempt them first and then ask you to check my work. For each problem, briefly indicate which specific skill it is testing.

Why it works: targeted practice on a specific skill at the right difficulty level is the most effective use of study time. Labelling what each problem tests helps you understand the curriculum structure.

Prompt 13: Exam Question Practice

Create a realistic exam question on [topic] similar to what would appear in [exam: e.g. GCSE Maths, A-Level Maths, university calculus, SAT]. Include a mark scheme after I attempt it. When I submit my answer, mark it using the mark scheme, explain any marks I lost, and tell me what a full-marks answer would include.

Why it works: working with mark schemes teaches you how examiners think. Understanding what earns marks versus what is just working towards an answer is a skill that significantly improves exam performance. Treat the generated mark scheme as approximate and check it against a real past paper where you can.

Prompt 14: Spaced Repetition Quiz

I want to test my recall of [topic or list of topics]. Ask me one question at a time. Wait for my answer before asking the next. If I get it right, move to a harder question on the same topic. If I get it wrong, explain the correct answer, then ask me a similar but slightly easier version to consolidate the concept. Keep going until I tell you to stop.

Why it works: this mimics the spaced repetition and adaptive difficulty that makes flashcard apps effective, but applied to maths problems rather than factual recall.

Prompt 15: Topic Summary for Revision

Create a concise revision summary for [topic] at [level]. The summary should include: (1) the key concepts and definitions in plain language, (2) the essential formulas or rules with a one-line explanation of each, (3) the most common question types and the method to use for each, (4) the three most important things to remember on exam day, and (5) one example of a common mistake and how to avoid it.

Why it works: a well-structured revision summary covers the same ground a good teacher covers in a pre-exam review lesson. Having it generated on demand means you can create one for any topic you need. Check every formula against your own notes before you revise from it.

Never revise from a generated formula sheet without checking it. A single wrong formula memorised in April is expensive in June, and it is exactly the kind of slip that arrives looking completely confident.

These prompts go beyond topic knowledge to build the deeper reasoning skills that make you genuinely good at maths rather than merely able to pass the next test.

Prompt 16: Understand the Problem Before Solving It

Here is a maths problem I need to solve: [paste problem]. Before we calculate anything, help me understand the problem fully. What is it actually asking? What information am I given and what do I need to find? Are there any constraints? What would a sensible estimate of the answer look like? Only after I understand the problem will we start solving it.

Why it works: rushing into calculation before understanding the problem is the root cause of most wrong answers. This prompt builds the professional habit of fully framing a problem before attempting a solution. The estimate step also gives you a way to catch an implausible answer later.

Prompt 17: Alternative Methods

Here is a problem I have solved using [method]: [paste problem and solution]. Show me at least two other methods that could solve the same problem. Explain the advantages and disadvantages of each approach and tell me which method is most appropriate for which type of situation.

Why it works: knowing multiple methods for the same problem is a hallmark of deep mathematical understanding. It also provides a built-in way to check your answers by solving the problem twice using different approaches, which is the most reliable way to catch an error.

Prompt 18: Build Intuition with Examples

I want to build better intuition for [concept: e.g. what makes a function continuous, when a matrix is invertible, what standard deviation actually measures]. Give me five numerical examples that together build my intuition: start with the simplest possible case, add one layer of complexity at a time, and explain what each example is designed to make me notice.

Why it works: mathematical intuition is built through carefully chosen examples, not through definition-reading. Sequencing examples to build incrementally is something good teachers do explicitly but most resources do not.

Prompt 19: Explain a Concept to Test Understanding

I am going to explain [concept] to you as if you are a student who has not learned it yet. Please listen to my explanation, then tell me: what did I get right, what did I miss or oversimplify, and what misconceptions might someone develop from my explanation? Here is my explanation: [write your explanation of the concept].

Why it works: this is the Feynman technique implemented with AI. Trying to explain something in your own words exposes exactly what you do and do not understand, and getting targeted feedback closes those gaps. It is also the one prompt here where the model cannot get the arithmetic wrong, because there is none.

Prompt 20: Build a Learning Roadmap

I want to learn [specific goal: e.g. calculus from scratch, enough statistics for data science, A-Level Maths in 6 months, university-level linear algebra]. My current level is: [describe what you already know]. Build me a structured learning roadmap that includes: (1) the prerequisite topics I need to master first, (2) the sequence of main topics in the order I should learn them, (3) why that order matters - what each topic unlocks, (4) rough time estimates for each stage, and (5) how I will know when I am ready to move on.

Why it works: one of the biggest barriers to self-directed maths learning is not knowing what to learn next. A dependency-ordered roadmap removes that uncertainty and prevents the frustrating experience of hitting a wall because a prerequisite was skipped.

Two habits matter more than model choice here. The first is doing the calculation yourself and asking only for the method, which removes the accuracy problem completely. The second is asking a second model to check a worked solution when the answer matters, because they tend to make different slips rather than the same one. Claude Sonnet 5 tends to hold the hint-only instruction rather than drifting into solving, which matters for prompt 7, while GPT-5.2 is quicker at generating practice sets.

If you are teaching rather than learning, our ChatGPT prompts for teachers cover the planning side, and prompt 12 is useful for producing thirty variations of one question so a class is not all working the same problem. For prompts worth memorising across every subject, see our best ChatGPT prompts collection. Chat Smith is free to try.

And the one thing worth repeating from the top of the page: verify the numbers, never submit the output, and attempt the problem first. Everything on this list works if you keep doing the maths yourself, and none of it works if you stop.