One TMUA Score, Two Admissions Systems

The Test of Mathematics for University Admission (TMUA) is a single admissions test whose one score, reported on a 1.0 to 9.0 scale, is now read by two universities with different decision structures. Oxford’s Mathematical Institute uses it to shortlist around three applicants per place. Cambridge Colleges use it to identify applicants unlikely to be competitive, and the Faculty of Mathematics still interviews 75 to 80 per cent of them. Neither publishes a cut-off.

Until this cycle the question did not arise. Oxford ran its own Mathematics Admissions Test from 2007 to 2025 and Cambridge used the TMUA, so a candidate produced a different number for each university and there was nothing to compare. That has ended. The Mathematical Institute states that for the 2026 admissions round the MAT will not take place and all applicants for Maths, Computer Science or joint honours courses take the TMUA instead — the change we cover in detail at the MAT has been retired, and which the department sets out on its own TMUA page. From the 2027 entry cycle onwards, one score goes to both.

Two universities receiving the same number is not the same as two universities agreeing what it means. They do not, and both say so in writing on their own departmental pages. This page sets out what each publishes, what each declines to publish, and what follows for a candidate who has to choose between them.

One test, one number, one sitting

Before either university sees it, the score is fixed by the test administrator. UAT-UK, the collaboration between Imperial College London and the University of Cambridge that owns the test, reports the TMUA as a single overall figure on a scale running from 1.0 at the low end to 9.0 at the high end, to one decimal place. There are two papers — Applications of Mathematical Knowledge, and Mathematical Reasoning — and Oxford requires both. The two papers do not produce two results. They produce one.

The sitting is shared as well. Oxford’s applicant guide states that candidates who are also applying to other universities that use these tests need only take the relevant UAT-UK test once. Whatever you score on the morning you sit it is the number that reaches every institution on your UCAS form. There is no version of the score tuned for a particular university, no separate Oxford paper, and no opportunity to sit again in the same cycle for a second reading.

That matters more than it sounds. When two universities each ran their own test, a candidate could reasonably think of their performance as two separate events with two separate outcomes. It is now one event. The variation has moved entirely to the receiving end.

Source: UAT-UK TMUA Explanation of Results for the October 2025 and January 2026 sittings; University of Oxford course pages.

Who reads your score at Oxford, and who reads it at Cambridge

This is the difference that produces every other difference on this page, and it is the one the commercial guides ranking for TMUA score queries do not state. At Oxford, the decision your score feeds into is taken by the department. At Cambridge, it is taken by a College.

Oxford’s Mathematical Institute publishes its own admissions timeline, and the entry for the month after the test reads simply: Oxford makes shortlisting decisions in November 2026, with interviews in December 2026 and decisions sent to candidates in January 2027. The actor is the university. The department explains why the test exists at all in terms of its own capacity: it cannot interview every applicant in the time available, so it shortlists.

Cambridge describes something structurally different. Its guidance on how a decision is made states that all applications are assessed carefully by the Cambridge Colleges, and that College Admissions Tutors consider all the information available together before making any decisions. On admissions tests specifically, the university states that Colleges use the test results to help them decide whether to invite you to interview, and that if you are shortlisted your test score is looked at together with the rest of your application before the College decides whether to offer you a place. The College appears three times in two sentences. That is not an accident of phrasing.

Cambridge also publishes the full list of what a College weighs: academic record, school or college reference, personal statement, any submitted written work, how well you do in your written admissions assessment, contextual data and any extenuating circumstances, and performance at interview if you are interviewed. The test score is one named item on a list of seven. Oxford’s Mathematical Institute frames its own use more narrowly — the information about the test, together with all the details of your UCAS application and information about school background, decides who to shortlist.

Source: Oxford Mathematical Institute TMUA page; University of Cambridge Undergraduate Study, admission tests and application decisions pages.

Oxford builds a list; Cambridge removes a tail

Both universities publish what they are doing with the number, and the two statements point in opposite directions.

Oxford’s Mathematical Institute writes that it shortlists around three applicants per place to interview, and that the TMUA is what helps it decide who those three are. The score is being used to build something: a list, of a fixed size, from a much larger field.

The Cambridge Faculty of Mathematics writes the opposite. TMUA performance, it says, is considered alongside other components of an application and used to help identify candidates who are unlikely to be competitive in the admissions round, rather than to select only those who are most likely to receive offers. The score is being used to remove something: the part of the field that will not make it, so that the rest can be looked at properly.

The published proportions follow from that. Cambridge states that historically between 75 and 80 per cent of Mathematics applicants have been invited to interview, and that it expects to continue interviewing the substantial majority. Around 30 per cent of all applicants have ultimately received conditional offers in previous years. So at Cambridge the TMUA is the first of three gates and it is the widest of them: most applicants pass it, and the interview and the STEP condition do the rest of the work. Our guide to why Cambridge Maths requires both the TMUA and STEP covers that second stage.

At Oxford, the shortlist is the narrow point. A candidate who is not shortlisted in November is not interviewed in December and receives nothing in January. There is no later stage at which a strong interview can rescue a marginal test performance, because there is no interview.

What this means for a mid-range score is worth stating carefully, because it is an inference from published policy and not a published figure. Neither university says what a mid-range score does. But a screening instrument that lets three-quarters of the field through and a shortlisting instrument that admits three per place are not the same instrument, and the same number cannot sensibly be read the same way through both. A score that sits in the middle of the national distribution is doing very different work in a process designed to keep most people in than in one designed to reduce a field to a fixed list.

At Cambridge, the College you name changes what sits beside your score

Because the decision is a College decision, the requirements around the score are College requirements, and they are not uniform. This is checkable rather than folkloric, and Cambridge publishes the cases.

The clearest is Computer Science. Every applicant takes the TMUA — but the course page adds that if you are applying to Peterhouse or Trinity you will also need to take the Computer Sciences Aptitude Test. Two Colleges, one extra test. A candidate with a given TMUA score who names Trinity is presenting that score alongside a second piece of subject evidence; the identical candidate who names a different College is presenting it alone.

Mathematics shows the same pattern at the offer stage. Cambridge publishes a flexible offer scheme for 2027 entry under which some Colleges will accept a slightly lower A-level profile in exchange for stronger STEP performance, and it names them: Downing, Emmanuel, Girton, Jesus, Newnham, Robinson, Sidney Sussex and Trinity Hall. Eight Colleges in, the rest out. The Faculty also notes that some Colleges run written assessments as part of their interview process, so the evidence base a College is working from at the point of decision differs between Colleges.

Then there is the pool. Cambridge explains that a College impressed by an application but without a place may put that application forward to be considered by other Colleges, because Colleges would rather admit a strong applicant who applied elsewhere than a weaker one who applied to them. In practice this means one TMUA score can be read by two separate admissions teams in the same cycle, at different moments and against different remaining places — the mechanism we set out in our explainer on how the Cambridge winter pool works.

Oxford has college-level machinery too: applicants may be interviewed at more than one college, and additional interviews are arranged after the initial ones. But that machinery operates after the shortlisting gate, not at it. The score has already done its decisive work by the time a second Oxford college looks at anyone.

What neither university publishes

The question every candidate asks is what score they need. The honest answer is that neither university answers it, and saying so plainly is more useful than an estimate dressed as a figure.

Oxford’s Mathematical Institute states that there is no pass mark for the TMUA. Cambridge states that there is no pass or fail for the TMUA and that candidates should aim to do the best they can. Neither publishes a minimum, a typical successful score, a banding, or a threshold on its admissions pages for any of the courses that require the test. Oxford has an additional reason for silence that will not apply forever: the 2027 entry cycle is its first with this test, so it has no historical TMUA data of its own to publish even if it wanted to.

The only published interpretation of where a number sits comes from the test administrator rather than from either university. UAT-UK states that the scale has been designed so that typical candidates score around 4.5, that approximately 10 per cent of candidates achieve scores higher than 7.0, and that scores are capped at 1.0 at the bottom and 9.0 at the top. That tells you your position in the national field. It does not tell you what any admissions tutor will do with that position, and UAT-UK does not claim it does.

Anything more specific that you find is somebody’s estimate. It is worth knowing how thin the published record is treated even by pages that rank well for this exact question: one of the higher-ranking commercial guides tells readers that no specific details are provided regarding how Oxford and Cambridge use the results, and speculates about what is likely. That is simply not the case. Both publish their policy, on their departmental admissions pages, in plain terms — it is just not on the central prospectus page where a guide writer would look first. A page that reports no policy exists and then supplies a threshold anyway has produced the number from somewhere other than the university.

When your score exists, and when each system acts on it

The calendar decides whether you can react to the number at all, and both universities publish theirs. The dates below are for entry in 2027, which is the first cycle in which the two universities read the same test.

Source: Oxford Mathematical Institute TMUA key dates; University of Cambridge TMUA page; UAT-UK test results page; Cambridge Faculty of Mathematics how to apply.

Two consequences follow from the ordering. The first is that you commit to a university before you know your score: the UCAS deadline falls on 15 October, inside the test window itself, and results arrive about a month later. Nobody chooses between Oxford and Cambridge on the basis of a TMUA result, because the choice is made first. If that choice is still open, our comparison of how to choose between Oxford and Cambridge works through the other grounds.

The second is that the waiting period differs in character. Results released around mid-November land in the same month Oxford makes its shortlisting decisions, so an Oxford candidate typically knows the number at roughly the moment it is being used, with no ability to act on it. A Cambridge candidate is waiting on a College whose timing is its own: Cambridge states plainly that Colleges make decisions about applications at different rates, and that hearing nothing while a friend at another College has heard means nothing about the outcome.

Reading your own score when you can only apply to one

A candidate applies to Oxford or to Cambridge in a given cycle, not both. That makes the difference in reading conventions a real decision input rather than a curiosity, and there are three things worth taking from it.

First, the two universities give opposite advice about preparing for the identical test. Cambridge states that the focus is on mathematical thinking, that it uses mathematics you will already have learned at school, and that there is no need to take a preparation course or to do much preparation beyond familiarising yourself with the format. The Oxford Mathematical Institute runs a weekly online livestream offering admissions support and problem-solving strategies. Neither is wrong; they reflect what each institution does with the result. A screening instrument rewards familiarity and composure. A shortlisting instrument rewards the marginal mark. Our TMUA preparation guide sets out the format in full.

Second, the strength of the rest of your application is worth more at Cambridge, and worth it earlier. Seven named elements go into a College decision, and the great majority of Mathematics applicants reach the interview at which several of them are weighed. At Oxford the shortlist is built from the test, the UCAS application and school-background information, and everything else happens after a gate that admits three per place. A candidate whose real strength is conversational mathematics rather than timed multiple choice is choosing between a system that will probably let them demonstrate it and one that may not.

Third, the academic profile around the score is severe at both and Cambridge publishes it. Among Mathematics students who took A levels and started in 2018, 2019 and 2023, 88 per cent of entrants achieved at least A*A*A*; for Computer Science the figure is 87 per cent. Whatever the TMUA is doing, it is doing it among applicants of that profile, which is the honest context for any score you are holding. It is also why the national distribution figures published by UAT-UK overstate your standing: the typical TMUA candidate is not the typical Oxford or Cambridge applicant.

What none of this supports is a number. There is no threshold to hit, because neither university publishes one, and a page that supplies one is guessing on your behalf about a decision made by people who have declined to describe it that precisely. What the published record does support is a clearer question than "is my score good enough": ask instead what the score is being asked to do at the university you have chosen, and what is left to do the rest of the work.

Why Families Choose Leading Tuition for TMUA Preparation

We are rated Excellent on Trustpilot, 4.8 out of 5 from 54 reviews, and 91% of our students achieve their desired grades. Those are practice-wide figures across every level we teach, not a TMUA-specific cohort — we do not publish a per-test success rate, because we do not hold one we could evidence. What we do is plan preparation around the October sitting and around which of the two systems the candidate has actually applied to, since the same score is not being asked to do the same job in both.

Frequently Asked Questions