COMPARISON
Where Mathiter fits next to the tools you already run
We are not going to tell you your tools are bad. You chose them for reasons that are still true, and three of the four here do things Mathiter cannot. This page exists because the fastest way to explain what Mathiter is, is to put it next to what it isn't.
Every competitive claim below links the vendor's own published page and says when it was read. If one has gone out of date, or we have read it wrong, email marketing@mathiter.com and we will change it.
The stack, as it actually is
Most international-school maths departments run four things at once. An LMS holds course structure and grades. A practice tool gives standards-mapped drill with a score. A free video library sits alongside it as a bookmark. A normed instrument measures growth two or three times a year.
All four answer the same question: did the student get it right, and how often. None of them has a field anywhere in its data model for why a particular wrong answer was chosen.
Which means only one line on that list is a replace decision at all: the practice-and-skill-score tool. The LMS stays. The normed measure stays. The free library stays.
IXL
What IXL has that we do not
An enormous bank across K–12 and beyond. A Real-Time Diagnostic covering maths and ELA in roughly 45 minutes per subject, producing grade-level knowledge levels that update as the student works. The most mature integrations on this page — Canvas with SSO and optional admin-enabled grade passback, plus Google Classroom, Clever, ClassLink and Schoology. And a Strong rating on Evidence for ESSA, which Mathiter does not have and will not have until a pilot generates one. On integration, we are not close.
Source: ixl.com/diagnostic · evidenceforessa.org/program/ixl-math (one cluster RCT, grades 3–5, four schools, n=545, +0.13 on Star Math) — read August 2026.
What happens after a wrong answer
IXL publishes how its SmartScore behaves. Below 70, correct answers move the score a lot and wrong answers cost little. At 80, a learner is considered proficient. In the 90–100 Challenge Zone a correct answer is worth 1–2 points and a wrong one costs 3–8, and a student may need as many as ten correct in a row to finish. IXL's own guidance for a missed question is that the student should read the explanation to get the key steps and strategies for the next one.
That is a worked account of the correct method — the mathematics, restated. Mathiter's response to the same wrong answer begins from the wrong idea that produced it, and refuses to state the answer until the student gets there. Neither of those sentences is an opinion: one is IXL's own PDF, the other is a description of our own product.
Source: IXL SmartScore Guide (PDF) — read August 2026.
What “IB coverage” means
IXL publishes IB Diploma Programme skill plans. IXL's own footer on those pages states the alignments are provided by IXL and are not affiliated with, sponsored by, reviewed, approved or endorsed by the International Baccalaureate. They are mappings of existing US-grade-level IXL skills onto IB syllabus points, and several named points come back with no skills mapped at all — Analysis & Approaches SL 1.6 (simple deductive proof, and the notation for equality versus identity), SL 5.5 (anti-differentiation with a boundary condition), SL 5.10 (integration by inspection or substitution), among others.
The pattern is not random. What survives the mapping is anything expressible as a repeatable single-skill drill. What comes back empty is proof, calculator-dependent technique and substitution. And neither the US nor the UK IXL site lists a Cambridge International plan — no 0580, no 9709, no 9231.
Source: IXL IB Diploma Programme skill plans · ixl.com/skill-plans · uk.ixl.com/skill-plans — read August 2026.
Our own concession, immediately
Mathiter's IB and Cambridge lecture coverage is thin too: IB AA has 22 lectures against 114 standards, IB AI 30 against 119, IGCSE 0580 14 against 128, Further 9231 none at all. The claim here is about the standard tree, the item bank and the misconception layer reaching those syllabi — not about lecture parity. The matrix on the curriculum page prints every one of those numbers.
See the coverage matrix →Price
IXL does not publish school pricing; its schools pricing page offers a quote. Neither do we. We are not going to guess at theirs.
Source: IXL schools pricing — read August 2026.
Khan Academy
Concede this first, because all of it is true
Free, with no contract and no procurement. Fifteen years of library across every subject a school teaches. Parents trust it without being told to. And its SAT preparation is the College Board's own partner product — not an alignment claim, a partnership. Nothing on this page positions against Khan's SAT prep as though it were unofficial, because it is the official one and ours is not.
Source: College Board — Official Digital SAT Prep with Khan Academy — read August 2026.
The mastery signal counts errors; it cannot name one
Attempted is under 70% correct. Familiar is 70–99%. Proficient is 100%. Mastered adds a mixed-skill assessment. Every rung is an accuracy count. Two students sitting at Familiar on the same quadratics skill — one dropping a sign at the zero-product step, one mis-factoring when the leading coefficient isn't 1 — are the same pixel on the dashboard, and they need different lessons tomorrow.
Source: Khan Academy help centre — read August 2026.
The remedy is the same ladder regardless of which wrong option was chosen
A hint sequence and a related video. Nothing in Khan's published reporting inventory classifies an error type.
Source: Khan Academy help centre — read August 2026.
No IB, no Cambridge, no ACT
Khan's test-prep index lists SAT, MCAT and LSAT. For an American-curriculum school Khan's maths coverage is close to total; for a school running IB or Cambridge alongside, it covers the American half of the timetable. This one is checkable by the reader in thirty seconds, which is why it belongs here.
Source: Khan Academy help centre — read August 2026.
What is purchasable at an address outside the US
Khanmigo's free teacher tools reach nearly every country; Khan's own help centre states that paid Khanmigo subscriptions are only available in the United States at this time, and Khan Academy Districts — the route to student Khanmigo at scale — requires a 250-student minimum and rostering through Clever or ClassLink, though non-US partnerships do exist through local partner entities. That is a fact about distribution, not a criticism of an organisation that gives its work away.
Source: Khan Academy help centre — read August 2026.
Rostering-gated integrations systematically exclude international schools. That is why we build toward Google sign-in and a pasted URL rather than toward Clever.
Canvas — and the same argument for Google Classroom
This is not a comparison
Canvas is your system of record and it stays your system of record: course structure, modules, assignment records, the gradebook, the roster, communication. None of that moves.
What Mathiter does not have, listed before anything else
No LTI 1.3, no LTI 1.1, no Deep Linking, no AGS grade passback, no NRPS roster pull, no Common Cartridge, no OneRoster, no Clever, no ClassLink, no Google Classroom integration. Today Mathiter is a link in a Canvas module, or a post in a Classroom stream. We will not write “integrates with Canvas” while that is a hyperlink.
The precedent is already in your building: Khan Academy has no first-party Canvas LTI either, and its 2025-26 LMS sync requires a Districts partnership, Clever or ClassLink rostering, a 250-licence minimum and a US district — structurally unavailable to an international school. Your department is already living, happily, with a major maths tool that has zero grade flow.
Source: Khan Academy — LMS sync requirements — read August 2026.
The one substantive thing worth saying about an LMS maths gradebook
For mathematics, New Quizzes reduces to two auto-markable types: Formula (randomised variables, student types a numerical answer, marked against a computed value) and Numeric (a number with a tolerance band). There is no algebraic equivalence check — 2x+1 and 1+2x are not the same answer — so expression-valued, interval-valued and show-your-working answers cannot be machine-marked at all. Matching and Categorization are all-or-nothing with no partial-credit option, which is precisely wrong for a question that asks a student to sort six functions into three families. MasteryPaths caps at three pathways per scored item and does not apply due dates to conditionally released items. And the data model records correct/incorrect, points, and optionally an outcome rollup: a distractor is a string with false next to it. There is nowhere in it to put a misconception, a trap type, a severity or an authored chain.
That is a gap by design, not by neglect. Canvas is a course-management system, not a diagnostic one.
Source: Canvas New Quizzes question types — read August 2026.
And the equation-input tax
The rich-content equation editor is click-driven or LaTeX. Neither is a thing you ask a Grade 9 student to do forty times. The third-party industry around this exists because of it.
Google Classroom — the same argument, shorter
If your gradebook is Google Classroom rather than Canvas, everything above holds and one thing improves: Google sign-in already works. Students use the Workspace account the school gives them, with no second password and no account creation. There is still no Classroom integration, no roster import and no assignment sync. At a year group of eighty-five, adding students by hand is an afternoon. That is the size of the work; it is not an integration and we will not describe it as one.
NWEA MAP and the measurement layer
What MAP is for, and keeps doing
A nationally normed, longitudinal growth measure, run two or three times a year, reported to leadership and to parents, comparable across cohorts and across years. Mathiter does not replace it and does not try to.
Where it stops
It gives you a RIT score and a growth trajectory. It does not tell you which misconception is in the room this Tuesday, and it does not run on Tuesday.
What Mathiter's placement adds next to it
The same kind of object — an ability estimate with a stated confidence range and a percentile — but free, uncapped, runnable inside one forty-minute period in any week of the year, and attached to a standard tree that names the specific leaf standards the estimate was built from. Then the misconception layer underneath it, which a growth measure has no equivalent of by design.
The honest asymmetry
MAP is externally normed and validated. Mathiter's placement is calibrated from its own response data. Different instruments, different warrants. If your leadership reports a normed number, keep reporting the normed number.
Eedi — the one tool built on the same premise
Eedi is built on the same idea we are: multiple-choice diagnostic questions where each wrong answer is authored to reveal a specific named misconception, across roughly 2,000 maths skills, with machine-learning work done alongside Microsoft Research and an EEF-registered evaluation. If your department already runs Eedi, we have less to tell you than we have to tell most people, and we would rather say that than let you find it out.
Source: Eedi — what is Eedi — read August 2026.
Three specific separations
Eedi's tree is UK National Curriculum and US skills — not IB AA/AI, not Cambridge 9709, not the AP suite. Eedi diagnoses and then routes onward, to a lesson or a tutor; Mathiter runs an authored Socratic chain with per-step hints that withholds the answer. And Eedi has no exam simulation and no IRT placement returning an ability estimate with a confidence range.
Other tools we get asked about
Knowing the shelf is itself a kind of credibility, so here is ours.
| Tool | What it is |
|---|---|
| Kognity | IB-native and genuinely deep — claims full AA and AI coverage at SL and HL with IA guidance; teacher access is free, student access is quote-only. If your school is IB-only, Kognity owns content depth we do not have yet.source |
| Mathspace | The closest per-step model to ours — the student works line by line and gets a hint at each line. Publishes IB textbook content by topic.source |
| ALEKS | The most serious adaptive engine on this list, built on knowledge-space theory with periodic Knowledge Checks. Coverage runs to college precalculus — it stops below AP, IB and A Level.source |
| Desmos / Amplify Classroom | A task-and-discussion tool, not a rival. It competes for lesson minutes, not for diagnosis. We embed Desmos inside the SAT simulator.source |
| MyiMaths | The international sibling of MyMaths, described by OUP as covering IB, Cambridge International, Edexcel and OxfordAQA for ages 11–18.source |
| DeltaMath | Free tier plus paid; integrates with Google Classroom, Canvas, Schoology, ClassLink and Clever, with availability depending on licence level.source |
What Mathiter does not replace
Your LMS and your gradebook
No integration, no write-back. Canvas or Classroom stays exactly where it is.
Your SIS and your roster
No import, no sync, and no class or section object that maps to your timetable.
Your normed growth measure
MAP stays. Our placement is a different instrument answering a different question.
Your curriculum or your textbook
We hold a standard tree and a question bank. We do not sequence your course.
Khan Academy
It is free, it is the official SAT partner, and there is no reason to drop it.
Your exam board
No accreditation, no credit, no transcript.
Free
Khan is. Our placement test is free and uncapped; the rest is not.
Your teacher
Everything here produces a diagnosis. A person still has to teach from it.
The only line where a replace decision genuinely exists is the practice-and-skill-score tool. That is the honest scope of the question, and we would rather say so than imply more.
The way to settle this is one class.
Nothing here asks you to drop anything. Run one class on Mathiter alongside what you already have, for a quarter, free, and compare the two on your own students rather than on our page.