| Component | Max | 4 | 5 | 6 | 7 |
|---|---|---|---|---|---|
| Overall | 100 | 40 | 52 | 64 | 76 |
HL Paper 1
Most questions were answered reasonably, with the triangle and sector question going best, but basic vector and complex number ideas, transformations and questions needing real thinking caused trouble.
Found difficult
- hypothesis testing
- complex numbers
- transformations of functions
- points of inflexion
- vector equation of a line
- phase portraits
- problem solving
Well prepared
- sine and cosine rules
- arc length and sector area
- normal distribution
- poisson distribution
- basic differentiation
| Q | Topic | How it went | Where marks were lost |
|---|---|---|---|
| 1 | t-test assumptions and hypotheses | Most struggled | could not state an assumption for the t-test; hypotheses without the population mean or written about samplesTip: Write hypotheses about the population mean and know that a t-test assumes a normal population. |
| 2 | arithmetic sequence | Mixed | misread the information, giving wrong common difference and first termTip: Read the context carefully to decide which value is the first term before finding the difference. |
| 3 | normal distribution and inverse normal | Answered well | could not use the inverse normal functionTip: Practise inverse normal on the GDC for finding a value from a given probability. |
| 4 | cosine rule and sector area | Answered well | applied the cosine rule to the wrong angle; used arc length instead of sector areaTip: Check whether the question asks for a length along the arc or the area of the sector. |
| 5 | multiplying complex numbers and equating parts | Most struggled | dropped the middle term when multiplying; did not then select the real component; did not equate real and imaginary partsTip: Expand complex products fully, then equate real parts and imaginary parts separately. |
| 6 | inverse variation model and its graph | Mixed | graph not through the given points; few correct final partsTip: Make sure your sketch passes through every point the question specifies. |
| 7 | optimization and solving an inequality | Mixed | found the x value but not the minimum value; found the critical values but not the interval between themTip: Answer exactly what is asked: the minimum value, or the interval, not just the boundary points. |
| 8 | transformations of graphs | Most struggled | transformations poorly understood; part (b) beyond even strong candidatesTip: Practise applying each transformation to key points and sketching the result step by step. |
| 9 | Poisson then binomial probability | Mixed | did not see a binomial distribution was needed in part (b)Tip: When counting how many of several periods meet a condition, think binomial. |
| 10 | differentiation and point of inflexion | Most struggled | quotient rule with an extra term; poor algebra; set the first derivative to zero for an inflexionTip: A point of inflexion needs the second derivative to be zero and change sign. |
| 11 | linear graph through the origin with known gradient | Mixed | could not sketch a line through the origin with a found gradient |
| 12 | simple and connected graphs and adjacency matrix powers | Mixed | did not know the meaning of a simple graph; used one matrix power instead of the sum of powersTip: To count walks of up to n steps, add the adjacency matrix powers from 1 to n. |
| 13 | vector motion with speed and least distance between paths | Most struggled | ignored the given speed and the direction vector length; could not minimise the distance expressionTip: Scale the direction vector to match the given speed, then minimise the distance with the GDC. |
| 14 | phase portrait and coupled rates | Most struggled | found dx/dt and dy/dt at a point and stopped; very few correct solutions to part (b)Tip: Use the computed rates to describe the direction of motion in the phase plane. |
| 15 | volume of revolution about the y-axis with unknown limit | Most struggled | integrated pi y squared dx instead of pi x squared dy; could not solve for the unknown upper limit with the GDCTip: When the limit is unknown, set the integral equal to the target and solve graphically on the GDC. |
| 16 | transition matrix and a simple probability | Mixed | used the transition matrix for a plain probability questionTip: Use the simplest method that fits; not every probability needs the matrix. |
| 17 | calculus problem at the end of the paper | Most struggled | weak calculus understanding; ran out of timeTip: Budget time so the final question gets a genuine attempt. |
What to practise
- Read questions carefully and give final answers to three significant figures.
- Avoid rounding intermediate values.
- Cover the whole syllabus so no question is unfamiliar.
- Always show working so partial credit is possible.
HL Paper 2
Marks spread widely; time was not an issue, but the final parts of questions and the HL-only content late in the paper were found hard.
Found difficult
- interpreting results in context
- logarithms and indices
- hypothesis testing
- type i and type ii errors
- integration techniques
- matrices
Well prepared
- summary statistics
- chi-squared goodness of fit test
- use of technology
- modelling with functions
- basic differentiation
- eigenvalues and eigenvectors
| Q | Topic | How it went | Where marks were lost |
|---|---|---|---|
| 1 | box plots, grouped mean and chi-squared goodness of fit | Answered well | misread the box and whisker diagram; did not round up to a whole number in context; divided by the number of classes instead of total frequency; wrong degrees of freedom or conclusion inconsistent with the comparisonTip: Use the GDC with mid-interval values for a grouped mean and make your conclusion match your comparison. |
| 2 | rate of change, integrating dh/dt and percentage increase | Mixed | did not treat gradient as an instantaneous rate; forgot the constant of integration; could not express a factor of 1.2 as a percentage increase; used dh/dt in place of hTip: To get h from dh/dt, integrate and use the initial condition to find the constant. |
| 3 | quadratic bridge model and comparing with regression | Mixed | gave coordinates instead of a parameter; rewrote the given equation rather than forming a new one; solved dy/dx = 0 when something else was asked; compared model accuracy instead of the specified valuesTip: Read exactly which quantities the comparison should be made between before answering. |
| 4 | composite functions and logarithm laws in a points model | Mixed | could not simplify with log laws; missed that points are earned every day of the weekTip: Practise rewriting composite exponential and log expressions into a required form by hand. |
| 5 | sampling, binomial test for a proportion, type II error and chi-squared independence | Most struggled | wrong inequality in the alternative hypothesis; used P(X = 54) or the wrong tail; no hypotheses written for an unscaffolded test; did not combine cells with expected frequency below 5Tip: At HL every test needs hypotheses, a p-value and a reasoned conclusion; merge small expected cells first. |
| 6 | projectile motion by integrating velocity components | Most struggled | did not integrate with respect to t; treated the constant angle as a variable; divided by t and lost the t = 0 solutionTip: Factorise rather than divide by a variable, then decide which root fits the context. |
| 7 | matrix form of a recurrence sequence and exact eigenvalues | Most struggled | arithmetic errors adding terms; verified with numbers instead of matrix algebra; did not use an earlier result; did not use the quadratic formula for exact valuesTip: Prove general results with algebraic matrices, and use the quadratic formula when exact answers are requested. |
What to practise
- Teach every topic and keep HL algebra skills alive: log laws, algebraic matrices, calculus with parameters, exact quadratic solutions.
- Do statistical calculations such as means on the GDC.
- Train students to choose the right test and present hypotheses, p-value and conclusion.
- Practise critical regions and type I and II error probabilities.
- Check answers for reasonableness, such as a p-value never exceeding 1.
HL Paper 3
The paper was accessible, and errors came mainly from uncovered syllabus parts, careless reading and missing the links between parts of each question.
Found difficult
- travelling salesman problem
- separable differential equations
- reading the question
Well prepared
- travelling salesman problem
- modelling with functions
- logarithms and indices
- logistic models
| Q | Topic | How it went | Where marks were lost |
|---|---|---|---|
| 1 | graph algorithms for bounds and sums of random variables in a theme park | Mixed | repeated vertices or did not return to the start in nearest neighbour; listed vertices instead of the order of edges; did not understand the deleted vertex bound or best upper bound; squared given variances before addingTip: Understand why each algorithm gives an upper or lower bound, and give the output in the requested form. |
| 2 | logistic model, separation of variables and comparing models | Mixed | wrong derivative of a constant log term; could not separate variables or missed the link to the previous part; did not comment on both models; confused by swapped roles of the variables in a given formulaTip: Look for how each part's result is reused later, especially in long paper 3 questions. |
What to practise
- Give students plenty of long paper 3 style questions so they expect parts to build on one another.
- Teach the reasoning behind algorithms and bounds, not only the procedures.
- If a part cannot be done, suggest a sensible value and carry it forward.
Internal assessment
What moderators raised about the exploration this session, criterion by criterion.
- A. unclear aims, unexplained abbreviations and game contexts, and long descriptions of software use
- B. unlabelled graphs, undefined variables and representations added for completeness rather than insight
- C. mechanical use of textbook HL techniques or software printouts without adapting the mathematics
- D. results restated or described instead of evaluating methods, assumptions and alternatives
- E. procedural calculations with reliance on tools and little evidence of understanding
- E. test assumptions unchecked, chi-squared used to test regression fit and high order polynomials chosen for prediction
About this summary
After every session the IB sends schools a subject report: how candidates did on each paper and question, and what teachers should work on. This page is my reading of the Maths AI HL report for November 2025, in my own words. No text from the report or from the exam questions is reproduced, and it is not endorsed by the International Baccalaureate.
Session codes: TZ is the time zone variant of the papers. Corrections: pietro@pietromeloni.com.
Frequently asked questions
Which questions were hardest in Maths AI HL November 2025?
+
The questions most candidates struggled with were on t-test assumptions and hypotheses, multiplying complex numbers and equating parts, transformations of graphs and differentiation and point of inflexion. The tables above list every question discussed in the report, paper by paper.
What did candidates find difficult in November 2025?
+
Across the papers the examiners listed hypothesis testing, complex numbers, transformations of functions, points of inflexion, vector equation of a line, phase portraits, problem solving and interpreting results in context, among others.
What were the grade boundaries for Maths AI HL November 2025?
+
Maths AI HL: a 7 from 76 out of 100 overall in November 2025. The component boundaries are in the table on this page.
Is this the official IB subject report?
+
No. It is my summary of the report the IB issues to schools after each session, rewritten in my own words with a teacher's reading of it. No text from the report or from exam questions is reproduced. Your school's IB coordinator can access the original.