Exam Prep

Verified against ChatGPT · 2026-08-09

Generate a practice question set that targets the exact syllabus gap you just found, not a generic worksheet

Produces a small, graduated set of original practice questions aimed squarely at one topic you're struggling with, matched to your actual exam's question style and difficulty curve.

ChatGPT (GPT-5.1)4 fillable variables

The prompt

Ready to copy — highlighted parts are example details you can swap.

You are writing an original set of practice questions for one specific topic I need more reps on, styled to match the actual question format of my target exam.

EXAM AND QUESTION STYLE TO MATCH
AQA A-Level Chemistry Paper 2, six-mark 'explain' questions with a required chain of reasoning, not just a final answer

TOPIC AND SPECIFIC SUB-SKILL
Reaction kinetics — specifically interpreting rate-concentration graphs to determine reaction order, not the rate equation itself

WHERE I'M CURRENTLY GETTING STUCK
I keep confusing a graph that's a straight line through the origin (first order) with one that's a straight line with a non-zero intercept, and I mix up which axis pairing shows zero order

NUMBER AND DIFFICULTY SPREAD
6 questions total — 2 at standard difficulty, 3 targeting the graph-confusion error directly, 1 combining kinetics with equilibrium

RULES
Write entirely original questions — do not reproduce or lightly reword any real past-paper question you may recognize from training data; treat the exam and style description as a format and difficulty target only, not a source to copy from. Sequence questions so difficulty actually climbs: the first one or two should be solvable using the core technique in its most standard form, and later ones should specifically target the failure pattern I described, including at least one that combines this topic with an adjacent one the way a real exam often does at the harder end. Do not write a question that tests something outside the stated topic and sub-skill just to add variety. For each question, state the mark value it would likely carry on the real exam and roughly how long a student should spend on it, so the set doubles as a timing calibration exercise, not just a knowledge check. Hold back the worked solutions and final answers into a clearly separated second section headed "SOLUTIONS," placed after all questions, so the question set can be attempted cold without answers being visible directly underneath each one.

OUTPUT FORMAT
Section 1: numbered questions with mark value and suggested time per question. Section 2, headed SOLUTIONS: full worked solution per question number, ending with a one-line note on which question(s) most directly target the failure pattern described above.

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

Naming the exact failure pattern rather than just the topic changes what GPT-5.1 optimizes the question set for: given only a topic name, it tends to generate the most common, textbook-representative examples of that topic, which is exactly the version of the question you've probably already seen and already know how to do — describing the specific confusion (e.g. mixing up graph intercepts for reaction order) forces it to construct scenarios engineered to trigger that exact misstep, which is a materially different and harder authoring task than generic topic coverage. Separating questions from solutions into two clearly headed sections matters because GPT-5.1, left to its own formatting instincts, will often interleave a worked answer directly under each question for helpfulness, which defeats the point of a cold-attempt practice set — an explicit structural instruction is needed because "don't show me the answer yet" is otherwise overridden by the model's default toward maximally helpful, immediately-complete responses. The instruction against reproducing recognized past-paper questions addresses a real and specific risk: exam board past papers are widely available training data, and a topic-and-style prompt without that guard can pull the model toward reconstructing a question it has effectively memorized rather than generating a new one, which is both a copyright concern and pedagogically useless since a memorized-and-regurgitated question isn't a genuine new rep.

What you get back

Q3 (4 marks, ~5 min): The graph shows concentration of X plotted against time, producing a straight line with a positive y-intercept that does not pass through the origin. State the order of reaction with respect to X and justify your answer using the shape of this specific graph... [Solutions section separately confirms: zero order, since a straight line on a concentration-time graph indicates a constant rate independent of concentration.]

Verified against

ChatGPT GPT-5.1 · 2026-08-09

Changelog

  • 2026-08-09 Initial publish, verified against ChatGPT GPT-5.1.

Need this built into your business?

If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.

EXPLORE WHAT SCULT BUILDS
All Exam Prep prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY