IB Maths AI Sample Papers

IB Maths Applications and Interpretation (AI) asks you to make mathematical choices in context. A strong answer does more than produce a graph or statistic: it defines the situation, uses technology deliberately and explains what the result means. The official specimen materials show how this work is assessed at SL and HL.

Official IB Maths AI specimen papers and markschemes

Open the official IB Maths AI specimen papers (PDF)

How to use IB Maths AI sample papers

The PDF includes AI Papers 1 and 2 for SL and HL, the additional Paper 3 for HL, and the matching markschemes. The current AI external assessment papers use technology. Work from the section for your level and follow the conditions printed on that paper.

The pack belongs to the Applications and Interpretation course first assessed in May 2021. It is a guide to assessment form and question style, not a forecast of a later examination. Students in a transition year should verify the syllabus and paper conditions with their school.

Check what the specimen pack covers

Locate the SL or HL material and the matching markscheme before starting. Treat each paper as evidence of the working style and assessment conditions, not as a prediction of a later examination.

Prepare your technology before timed practice

Calculator fluency is part of AI preparation, but speed alone isn’t the goal. Check angle mode, graph settings and statistical lists before a session. Know how to enter a regression, read a distribution, solve an equation numerically and store enough precision for later steps. Practise writing the mathematical setup beside the technology output so the method can be followed.

• Clear or label lists before entering a data set.
• Check the viewing window against the variables and units.
• Keep unrounded values for calculations that continue.
• Write down the feature, model or probability you asked the calculator to find.

Run a complete modelling cycle. Keep the four stages visible so a calculator mistake, weak model or unsupported conclusion can be diagnosed separately.

Frame the situation

Name the variables, their units and the quantity the question asks for. Decide which information matters. If an assumption is needed, state it plainly rather than hiding it inside a calculator command.

Choose and use a method

Select a model, statistical process or numerical technique that fits the information. Enter it carefully, then preserve the relevant output. The exam is testing the choice of mathematics as well as the button sequence.

Interpret the result

Return to the original setting. Include units, sensible accuracy and the group or time period being described. For a regression or probability result, say what the number means instead of repeating the calculator display.

Challenge the answer

Ask whether the value is plausible and whether the model is being used within a reasonable range. Comment on a limitation when the question calls for it. An answer can be numerically correct but weak if it ignores context, assumptions or the reliability of the data.

Turn the AI markscheme into a data review

For each lost mark, record the stage that failed: framing, method choice, calculator entry, interpretation, validation, units or accuracy. Review the earliest failure first. A wrong conclusion may come from a poor model or data entry rather than from the final sentence.

Repeat the skill with a new data set or context. Changing the story helps test whether you understand the method, while repeating the original numbers may only test memory. Mix statistics, functions, geometry and calculus once the individual processes are secure.

Build up to an AI exam attempt

Start untimed while you make the modelling steps explicit. Add short limits to calculator tasks and interpretation questions, then combine them in a longer section. During a full paper, reserve time to reread contextual conclusions and check that rounded answers still match the requested accuracy.

Common AI mistakes to watch for

• Giving calculator output without defining the model or variables.
• Using a regression outside the observed range without considering whether it remains reasonable.
• Reporting a number with no units or contextual meaning.
• Rounding too early and changing a later result.
• Assuming that access to technology removes the need to explain mathematical decisions.

For each AI modelling question: frame the problem, calculate, interpret the result and test whether the conclusion makes sense. Use the specimen pack to practise that complete cycle under the conditions for your own level.