Q: What does International Baccalaureate (IB) Math Syllabus: 2025/26 Parent & Student Guide cover? A: Understand the AA vs AI pathways, assessment formats, and planning milestones for the 2025/26 IB Mathematics cohorts.
TL;DR IB Maths now comes in two flavours - Analysis and Approaches (AA) for proof-heavy algebra + calculus, and Applications and Interpretation (AI) for statistics + modelling power-users. Both streams keep the same 20 % Internal Assessment (IA), but their exam papers test rather different skills. Choosing well depends on a student's post-IP trajectory: engineering & physics prefer AA, while data-centric fields welcome AI.
HL extends to designing data collection, non-linear regression, central limit theorem, confidence intervals, Type I/II error analysis, and transition-matrix/Markov-chain modelling.
Both courses expect confident use of technology: GDCs are mandatory in calculator papers, and the guides push spreadsheets/dynamic graphing software throughout the syllabus (tech fluency is part of the assessment philosophy).
3 | Assessment in detail
3.1 External papers
Paper
Time
Calculator?
Skills hit
P1 AA (SL/HL)
90/120 min
No
Pure algebra & calculus drill
P2 AA (SL/HL)
90/120 min
Yes
Longer modelling Qs \(\Rightarrow\) partial-fractions, DEs
P1 AI (SL/HL)
90/120 min
Yes
Data interpretation, tech commands
P2 AI (SL/HL)
90/120 min
Yes
Extended statistical modelling
P3 HL (both)
60 min
Yes
Investigation / proof write-ups
Calculator rule: Papers with technology columns marked “Yes” require an IB-approved GDC throughout (Assessment procedures). Paper 1 stays non-calculator for AA SL/HL to emphasise analytic methods.
Tables condensed from subject briefs above.
3.2 Internal Assessment (IA)
Individually chosen exploration of roughly 12-20 double-spaced pages (quality over length per IB guidance).
Marked on five criteria - Presentation, Mathematical communication, Personal engagement, Reflection, Use of mathematics.
Allocate about 10-15 guided hours across briefing, draft check-ins, and authenticity checks.
Top-band notes: Criterion C expects personal engagement that “drives the exploration forward”, while Criterion E (HL) calls for mathematics that is “precise” and shows “sophistication and rigour” (guide descriptors).
Weight 20 % across all four courses .
4 | Which route for which goal?
Future pathway
Recommended course
Why
Engineering, physics, pure math
AA HL
Requires rigorous calculus \(\int\), series \(\sum\), proof.
Economics, computer science*
AA SL / AI HL
Balance of calculus with statistics and algorithmic modelling.
Life science, psychology, business
AI SL / AI HL
Heavy inferential statistics and real-data modelling.
Some CS faculties (for example Cambridge Computer Science) still request AA HL when the course is theory-heavy; always check university matrices.
Who thrives where? AA learners tend to enjoy algebraic manipulation, formal proofs, and pattern-spotting without leaning heavily on tech. AI learners usually light up when modelling real-world datasets and are comfortable testing conjectures with GDCs, spreadsheets, and dynamic tools (course introductions).
Insight from RevisionDojo comparisonand Knowledgeum Academy guide .
5 | Command terms decoded
The IB glossary (AA/AI guides, appendix) pins down exactly what examiners expect:
Notation heads-up: Exam questions follow the IB ISO-based notation list (Σ, μ, σ, etc.). Make sure you can read those symbols even if your CAS defaults differ-there’s no notation sheet in the exam room.
Term
IB definition
Exam tip
Interpret
Use knowledge to recognise trends and draw conclusions from information.
Always connect the maths back to the context (e.g. what a gradient means physically).
Investigate
Make a systematic examination to establish facts and reach new conclusions.
Lay out steps, state findings, and comment on implications.
Justify
Give valid reasons or evidence to support an answer.
Quote theorems/working explicitly, not just the final value.
Show / Show that
Give steps in a derivation; “show that” expects you to reach the stated result without full proof formality.
Start from given expressions, keep algebra line-by-line, and finish with the target statement.
Sketch
Draw a labelled graph that captures the key features.
Mark intercepts/asymptotes; scale can be rough but features must be clear.
Solve
Obtain answers using algebraic, numerical or graphical methods.
State all solutions relevant to the domain (include extraneous checks).
Verify
Provide evidence that validates the result.
Substitute back or reproduce working to confirm equality.
“Hence find ...” still signals you must reuse a prior result; quote it explicitly before continuing.
6 | Core formulae you still must know
IB supplies a booklet, yet speed matters. Burn in:
Binomial coefficient (kn)=k!(n−k)!n!.
Normal PDF f(x)=σ2π1exp(−2σ2(x−μ)2)
Differential rules: dxd(ekx)=kekx
Definite-integral area: ∫abvdt=s.
Drill until recall time < 3 s each.
Formula booklet tip: every centre issues a clean IB Mathematics formula booklet in the exam. Download the latest version from the Programme Resource Centre and rehearse with that layout so nothing feels new on paper day.
ℹ️ Grade statistics: IB publishes annual grade distributions in the Diploma Programme statistical bulletin. Download the latest issue from ibo.org before quoting cohort means in applications or marketing material.
7 | Crafting a high-scoring IA
Pick a narrow, curious question e.g., “Modelling long-jumper trajectory with a quadratic fit”.
Collect primary data (tracker app, spreadsheet import).
Deploy at least HL-level techniques: piece-wise regression, r2, residual analysis.
Reflect - strength, limitations, future work.
Match the maths to your level: IB’s criterion E awards top marks only when the techniques are course-appropriate (HL needs rigour/sophistication; SL demands fully correct syllabus-level work).
Aim to show multiple representations (algebra, graphs, technology output) so criterion B stays at the top band.
7.1 | Modelling frameworks (AA vs AI) for 2026 cohorts
Dataset idea
Description
AA angle
AI angle
Learning analytics log
Daily logins vs assignment completion for a Sec 4 IP class (anonymised)
Optimise logistic model parameters, prove convergence of Newton iteration for best-fit root
Run moving-average smoothing, build exponential smoothing forecast, discuss tech-enabled interpretation
Projectile Tracker export
\(x,t\) and \(y,t\) data from Tracker video of a physics launch
Derive parametric equations, compare analytical vs regression trajectory, explore angle for max range
Use regression diagnostics to interpret residuals, simulate air resistance via spreadsheet macros
JC finance time-series
Monthly tuition revenue vs marketing spend (synthetic but realistic)
Formulate system of equations, explore AR(2) model justification with characteristic roots
Run regression with correlation matrices, produce dashboard explaining elasticity & prediction intervals
🧰 Action step: Curate a shared Drive “IA modelling lab” and store anonymised datasets there so supervisors can audit ethical use. Encourage students to remix the packs-AA learners justify algebraic structure, AI learners lean on tech-driven inference. Every submission must still cite the dataset pack as a secondary source.
Workflow tips
Kick-off workshop: Spend 60 minutes letting students interrogate each dataset. Have AA groups outline the algebraic backbone while AI groups storyboard the technology they will deploy.
CAS vs manual balance: Require AA explorers to demonstrate at least one hand-worked derivation (e.g., solving for logistic parameters) before verifying with CAS. For AI, annotate every graph/table with the specific CAS/technology command used.
Common reflection prompts: “What assumptions underpinned the model?”, “Where did the technology aid insight-and where could it mislead?”, “How would you collect higher-resolution primary data?”
ATL alignment: The guides lean on inquiry, collaboration, and technology-rich investigation; plan IA checkpoints that explicitly build those approaches to teaching and learning (ATL) skills.
📎 Instructors: Keep a moderation log noting which dataset variant each student picked and the level of mathematical sophistication exhibited. This guards against accidental topic overlap when multiple candidates lean on the shared pack.
Deliverable add-ons
Publish a one-pager IA slog (Summary of Learning & Output Grid) outlining question, data source, representational forms, and integrity declarations. Collect via Google Forms for easy moderation.
Package the three dataset packs + templates as a downloadable bundle (Google Drive/Notion) linked in the CTA so parents/students have instant scaffolding.
