Cambridge Engineering ESAT scores: what the official FOI data shows

The United Kingdom's Freedom of Information Act allows anyone to request information held by public bodies, and universities fall within its scope. In 2025, two public FOI requests received formal responses from the University of Cambridge's Information Compliance Office. The first, reference FOI-2025-956 (answered on 17 October 2025), set out ESAT score statistics for Engineering (UCAS code H100) in the 2025 admissions cycle across all 29 colleges: for each college and each module (Mathematics 1, Physics and Mathematics 2), it lists the minimum, mean and maximum score for applicants and for offer holders separately. The second, reference FOI-2025-982 (answered on 3 November 2025), split the same cycle's figures by fee status, comparing Home and international applicants and offer holders.

Both responses are formal documents issued by the University itself and are published on WhatDoTheyKnow, the public FOI archive, so anyone can verify the originals. Two points of scope matter. First, the responses contain score statistics, not admissions rules: nothing in them amounts to a required mark. Second, Cambridge notes in the documents that some applicants change college or course during the cycle, and the figures are grouped by application year. Every number on this page and in the college explorer is taken directly from these two originals, with no extrapolation of our own.

Take the 25 standard undergraduate colleges together (setting aside the four mature colleges, which only admit students aged 21 and over) and the applicant pool centres around 4.7 to 4.8: the mean of college averages is 4.84 in Mathematics 1, 4.72 in Physics and 4.77 in Mathematics 2. That lines up closely with the picture UAT-UK paints in its official Explanation of Results, which describes the typical candidate as scoring around 4.5, with roughly ten percent of candidates above 7.0. Cambridge Engineering applicants are somewhat stronger than the overall test-taking population, but not dramatically so.

Offer holders are a different world. The mean of college averages for offer holders is 6.30, 6.40 and 6.23 across the three modules, roughly one and a half points above the applicant pool. In other words, the journey from submitting an application to holding an offer spans about a point and a half per module on average. The official anchors help put that in perspective: the applicant pool centres near the typical candidate, while the offer-holder pool has shifted decisively towards the top band. For anyone preparing, the first actionable conclusion is that matching the typical candidate is not the goal; the reference group that matters is the offer holders.

FOI-2025-982 is the more consequential of the two documents for international candidates. It shows that in the 2025 cycle, Home applicants to Cambridge Engineering averaged 4.25, 4.32 and 4.24 across the three modules, while international applicants averaged 5.51, 5.19 and 5.38. Before offers even enter the picture, the international pool is already visibly stronger.

At the offer-holder level the gap widens further. Home offer holders averaged about 5.67 in Mathematics 1, 5.85 in Physics and 5.67 in Mathematics 2; international offer holders averaged 7.41, 7.20 and 7.21, which is 1.74, 1.35 and 1.54 higher respectively. Set against UAT-UK's statement that roughly ten percent of candidates score above 7.0, the average international offer holder sits inside the top tenth of the entire test-taking population.

One thing must be said plainly: these figures describe an outcome, not a rule. No Cambridge document states that international applicants must reach a higher score, and the gap most plausibly reflects self-selection in the international pool combined with the intensity of competition for places. But the practical implication for international candidates is unambiguous: the group that actually receives offers in your lane averages at that height. Using structured practice and full mock exams to push yourself into the top band is the part of the outcome that stays in your own hands.

Lay the 25 standard undergraduate colleges out module by module and the differences in offer-holder averages are plain to see. In Mathematics 1, Murray Edwards (7.08), Christ's (7.05) and Downing (6.94) sit at the top, with St Catharine's (5.60), Girton (5.79) and Emmanuel (5.81) at the other end. In Physics, Clare (7.34), Selwyn (7.06) and Trinity Hall (7.00) lead, while King's (5.63), St Catharine's (5.86) and Churchill (5.90) trail. In Mathematics 2, Robinson (6.99), Murray Edwards (6.95) and Clare (6.80) are highest, with St Catharine's (5.70), Girton (5.78) and Selwyn (5.79) lowest. Averaging the three modules, Murray Edwards (6.89), Trinity Hall (6.80) and Clare (6.77) top the table while St Catharine's (5.72), King's (5.75) and Churchill (5.93) close it, a spread of roughly 1.2 from first to last.

The four mature colleges (Hughes Hall, St Edmund's, Wolfson and Lucy Cavendish) admit only students aged 21 and over. Their applicant pools are not comparable with the standard colleges and their samples are tiny, and in places the minimum, mean and maximum coincide exactly, which suggests a single person behind the cell, so their averages should not be read alongside the rest. The complete figures for all 29 colleges, including the minimum and maximum behind every cell, are available in the interactive explorer.

Seeing a 1.2 spread from top to bottom, the obvious thought follows: would applying to a college with a lower offer-holder average improve the odds? Slow down, for at least four reasons. First, the FOI responses do not include the number of applicants or offers per college, and without volume an average tells you nothing about how competitive any college actually is. Second, this is a single cycle's snapshot; the popularity of individual colleges shifts every year, and last year's quiet choice can be this year's crowded one. Third, Cambridge's winter pool redistributes strong applicants between colleges, so a college's offer holders are not necessarily the people who originally applied there, which further shrinks any gain from playing the college game. Fourth, a lower average may simply mean that college was willing, that season, to make offers to applicants whose test scores were less extreme once interviews and the wider file were weighed, which is not the same thing as being easier.

The sounder reading treats the college spread as context: it tells you what each college's offer-holder cohort looked like, not where to place a tactical bet. Choose a college on its own merits, its size, atmosphere, accommodation and location, and spend the effort you control on pushing your score into the offer-holder band.

The most practical use of this data is to replace hearsay about what counts as a competitive score with a reference frame that has a source. If you are an international applicant, your target band should align with the international offer-holder averages: all three modules in the region of 7, which by the official description is the top tenth of all candidates. If you are a Home applicant, the offer-holder averages sit between 5.67 and 5.85, well above the applicant pool but far less extreme than the international lane.

Two cautions belong here. First, an average is the middle of a group, not an entry ticket: single-module scores as low as 3.1 appear among standard-college offer holders, and plenty of high scorers did not receive offers, because interviews and the rest of the file genuinely matter. Second, a target is only useful when it shapes training: use topic-by-topic practice to close specific gaps, full timed mocks to stabilise your performance under pressure, and error analysis to convert every session into incremental score. Keeping in mind that the average international offer holder scored around 7.4 says more about the required intensity of preparation than any motivational slogan could.

For this data to be used well, its boundaries need stating plainly. First, averages are not admission requirements: Cambridge has never published a required score in any form, and the FOI data describes a distribution of outcomes, not a rule. Second, the responses contain no per-college applicant counts, so none of the averages can be weighted, and figures for small colleges will swing considerably from year to year. Third, the four mature colleges have tiny samples and should not be compared with the standard colleges. Fourth, the University notes that some applicants change college or course mid-cycle and the figures are grouped by application year, so college attribution carries a little churn. Fifth, this is a single snapshot of one course in one cycle: module combinations differ between courses (Cambridge Engineering fixes Mathematics 1, Mathematics 2 and Physics), so extrapolating to other courses has no basis, and any cross-year use should state the cycle.

Within those boundaries, it remains the most solid public reference frame for Cambridge ESAT scores in existence: official documents, college-level granularity, applicants against offer holders, Home against international. We will track future FOI releases in later cycles and keep this page and the explorer up to date.