ESAT preparation guide

ESAT in October or January: what the official data can settle

This is the question ESAT candidates ask most often and the one where public information is least reliable. UAT-UK's 2025/26 technical report answers half of it: the score distributions are printed separately for the October 2025 and January 2026 events, module by module, and the annual report adds the candidate counts and year-on-year growth for each window. This page puts every sitting-level figure from those two documents side by side, shows the gap, and then explains what the gap actually represents in measurement terms. One conclusion belongs up front: both events were scaled with the same constants, so a difference in means is a difference in who sat the test.
Practise 60 ESAT papers (2 free)Full ESAT guide
Papers
60
Questions
2473
Free to try
2
Quick answer: what actually differs between the sittings
Is the choice yours
One sitting per cycle and no January resit; a Cambridge course with an October UCAS deadline means October
Is it the same scale
Yes. Both events used the same scaling constants, so scaled scores are comparable across events within the cycle
Where the distributions differ
October: median 4.5 and 90th percentile 7.0 on all five modules. January Mathematics 1: 4.1 and 5.4
Is one sitting easier
No basis for saying so. The scale already removes form difficulty, so the gap describes the cohorts rather than the papers

Every figure here comes from a printed table or sentence in UAT-UK's 2025/26 ESAT technical report and its 2025-26 annual report. Nothing is read off a chart image and nothing is derived across tables. FrontierVue is an independent preparation platform and is not affiliated with UAT-UK or any university.

In short
  • Constraints come before data: one sitting per admissions cycle, no resits in January, and a Cambridge course with an October UCAS deadline leaves October as the only option. This cycle 10,116 sat in October and 4,077 in January, 71% to 29%.
  • In October the median is 4.5 and the 90th percentile 7.0 on every single module, because the scale is pinned to the October ability distribution. That column is the reference frame by construction.
  • January sits lower and tighter: Mathematics 1 has a median of 4.1, a 90th percentile of 5.4 and an SD of 1.03, against 4.5, 7.0 and 1.59 in October. Physics and Mathematics 2 show the same shape.
  • This is not a difficulty gap. The same scaling constants were used for both events, and the report states this was done so that scaled scores would be comparable across events.
  • Module coverage is a hard constraint: only 16 candidates took Biology in January and the report prints NA across that entire row, while Chemistry had 332. For those two modules January offers almost no benchmark.
01

Both sittings, module by module

This is the table in the technical report that answers the sitting question most directly: five modules, each read three ways, with candidate counts, means, standard deviations and percentiles for the whole cycle, the October 2025 event and the January 2026 event. Note one thing before reading it: in the October column the median is 4.5 and the 90th percentile 7.0 on every module. That is not a coincidence, and the next section explains it.

10,116

Candidates who sat in October 2025, 71% of the ESAT cycle

4,077

Candidates who sat in January 2026, 29%, against 23% a cycle earlier

4.5 / 7.0

October median and 90th percentile on every module, where the scale is anchored

One set of constants

Shared by both events so that scaled scores are comparable across them

Whole cycleOctober 2025January 2026
Mathematics 1 (compulsory)14,193 · mean 4.58 (SD 1.47) · median 4.4 · 90th 6.710,116 · mean 4.73 (SD 1.59) · median 4.5 · 90th 7.04,077 · mean 4.19 (SD 1.03) · median 4.1 · 90th 5.4
Mathematics 212,113 · mean 4.40 (SD 1.63) · median 4.2 · 90th 6.58,082 · mean 4.73 (SD 1.66) · median 4.5 · 90th 7.04,031 · mean 3.75 (SD 1.33) · median 3.7 · 90th 5.4
Physics11,290 · mean 4.25 (SD 1.71) · median 4.1 · 90th 6.57,717 · mean 4.58 (SD 1.70) · median 4.5 · 90th 7.03,573 · mean 3.55 (SD 1.51) · median 3.4 · 90th 5.5
Chemistry3,146 · mean 4.61 (SD 1.74) · median 4.5 · 90th 7.02,814 · mean 4.72 (SD 1.72) · median 4.5 · 90th 7.0332 · mean 3.63 (SD 1.62) · median 3.5 · 90th 5.7 · max 8.9
Biology1,547 · mean 4.44 (SD 1.87) · median 4.5 · 90th 7.01,531 · mean 4.46 (SD 1.86) · median 4.5 · 90th 7.016 candidates · the whole row is marked NA, with no statistics published
02

Why October lands on 4.5 and 7.0 every time

A scaled score is not a raw score. After each testing window the items are calibrated with a Rasch model, raw scores are turned into ability estimates, and the ability estimate is then transformed onto the 1.0 to 9.0 scale. The anchors are taken from October: median ability maps to 4.5, the 90th-percentile ability maps to 7.0, a regression line through those two points gives the scaling constants, and the result is capped at both ends. January is not re-anchored; it inherits the same constants. So the October column is defined, and the January column is measured.

The ESAT 2025/26 scale

Pinned to the October 2025 ability distribution, then reused unchanged in January

Anchor 1: median ability maps to 4.5The 50th-percentile ability at the October event, module by module (Table 2)
  • Mathematics 1 ability 0.4049
  • Physics 0.3454 · Chemistry 0.2960
  • Biology 0.5970 · Mathematics 2 0.0159
Anchor 2: 90th-percentile ability maps to 7.0The second row of the same table, again October only
  • Mathematics 1 ability 2.7701
  • Physics 1.8273 · Chemistry 1.9269
  • Biology 1.6789 · Mathematics 2 1.6918
A line through the two anchors gives the constantsTable 3, with a different constant and multiplier per module
  • Mathematics 1: constant 4.0720, multiplier 1.0570
  • Physics 3.9173 / 1.6870 · Chemistry 4.0463 / 1.5329
  • Biology 3.1205 / 2.3107 · Mathematics 2 4.4763 / 1.4917
The same constants are applied in JanuaryThe stated purpose: to keep scaled scores comparable across events
  • Transformed scores are capped at 1.0 and 9.0
  • The scaling is re-fixed each cycle, so cycles are not comparable

Tap a branch to unfold

This is the paragraph most often got wrong, so read it before drawing a conclusion. A difference in means between October and January does not mean one paper was harder or softer. The scaled score has already had form difficulty removed: within a cycle, whichever form you draw and whichever event you sit, your score lands on the same ruler, and both events were transformed with the same constants. The report's own illustration is blunt: a candidate scoring 6.5 in Chemistry, in either January or October, has higher ability than one scoring 4.2. The gap between the two means therefore describes two differently composed groups of candidates. On composition the documents do give one quotable rule: a candidate may sit only once per admissions cycle, and applicants to a Cambridge course with an October UCAS deadline must use October, which the technical report names as the reason that event carries the higher volume. How much of the score gap that rule accounts for is not something the report quantifies, and this page does not guess.
03

The same score sits in a different place in each sitting

Read the percentiles side by side and the shape of the gap tells you more than its size. On Mathematics 1 the 25th percentile is 3.6 in both sittings, while the 75th is 5.6 against 4.7 and the 90th is 7.0 against 5.4. The bottom quartile of the two cohorts nearly coincides; it is the upper half that separates. Across the four modules with published January figures, January carries the smaller standard deviation, so its scores are more tightly bunched. Biology has no January figures at all.

Mathematics 1, 90th percentile

7.0 in October, 5.4 in January

25th percentile 3.6 in both; SD 1.59 against 1.03

Mathematics 2, 90th percentile

7.0 in October, 5.4 in January

25th percentile 3.6 against 2.9; SD 1.66 against 1.33

Physics, 90th percentile

7.0 in October, 5.5 in January

25th percentile 3.4 against 2.5; SD 1.70 against 1.51

Chemistry, 90th percentile

7.0 in October, 5.7 in January

The January maximum is 8.9, the only printed maximum in Table 5 below 9.0

The practical reading: the same score represents the same ability in both sittings, but occupies a different position within each cohort. A 5.4 on Mathematics 1 is exactly the 90th percentile in January; in October the same score sits below that event's 75th percentile of 5.6. If you are using these numbers to locate yourself, be clear about which column you are comparing against. For the whole-cycle distribution, see the official ESAT score statistics.
04

Two different rooms: what the annual report prints about each cohort

Beyond the scores, the annual report breaks bookings, deliveries, the UK and overseas split, access arrangements and bursaries down by sitting. That is the whole of what the official documents give you about how the two groups differ structurally. The reports do say why October is the larger window and why the entry grew, and both explanations are quoted below. Structure and scores are still separate things, and this page keeps them separate rather than turning one into the cause of the other.
Bookings against tests actually delivered
October: 10,502 ESAT bookings and 10,116 tests delivered, with ESAT making up 48.5% of that sitting. January: 4,279 bookings and 4,077 delivered, 33.4% of that sitting. Across the cycle ESAT drew 14,781 bookings and 14,193 delivered tests, 43.1% of everything UAT-UK delivered. The gap between the two columns is candidates who booked and did not sit.
The two windows are growing at very different rates
Against 2024-25, October ESAT deliveries rose +11% and January rose +47%. The split between the windows is moving too: last cycle 77% of 11,919 Mathematics 1 candidates sat in October and 23% in January; this cycle it is 71% and 29% of 14,193. The technical report attributes the overall growth to three things: additional courses using the tests, an increase in applications to the courses already using them, and a higher proportion of applicants registering for the tests.
UK and overseas bookings by sitting
On bookings, October ESAT drew 5423 candidates in the UK and 5079 overseas; January drew 2257 in the UK and 2022 overseas (printed here without thousands separators, as the annual report prints them). Both windows are close to an even split, so this dimension separates the two sittings far less than raw volume does. Note these are bookings, not delivered tests, a different basis from the row above.
Access arrangements and bursaries (all three tests in that window)
801 candidates booked with access arrangements in October, 3.7% of that sitting, against 652 in January, 5.1%. UK bursaries: 1300 issued in October, 11.2% of UK candidates, and 1079 in January, 15.4%. Both sets cover every candidate in that window across all three tests, not ESAT alone, so they cannot be quoted as ESAT proportions.
05

Using this to decide which window to sit

01Hard constraint first

Check which window your course accepts

The first constraint is not the score distribution, it is the booking rules, and those are printed in the documents themselves. The annual report is explicit: a test may be taken only once per admissions cycle, and resits were not permitted in the January sitting. Both windows were open to everyone except candidates applying to a Cambridge course with an October UCAS deadline, who could only use October. The technical report attributes October's higher volume to that same rule. Once that is settled, each course's own deadlines and requirements still come from its admissions page.

02Sample size

Look at your module combination

The most common combination this cycle is Mathematics 1 plus Mathematics 2 plus Physics, taken by 74% of candidates, and all three have thousands of candidates in both windows. Biology and Chemistry are different: January had 16 Biology candidates, with every statistic on that row printed as NA, and 332 in Chemistry. If your combination includes those modules, January gives you almost no benchmark and you will have to locate yourself against October or the whole cycle.

03Two test days

Weigh readiness against the calendar

The technical report gives this cycle's two windows: 9th and 10th October 2025, and 6th and 7th January 2026. The annual report adds that each window spans six days of testing with every test delivered across two of them, and that for test-security reasons candidates in China, Hong Kong or Macau could sit on only one day per sitting. Each of the five ESAT modules is 40 minutes and 27 multiple-choice items, taken consecutively. October frees you earlier but compresses revision into the summer and the start of term; January buys a few more months and spends them in the busiest stretch of the application year. No official figure decides this one; your own readiness does.

04£75 / £130

The fee is the same either way

A test costs £75 for candidates in the UK and Republic of Ireland and £130 everywhere else, identical in both windows, so cost is not a reason to pick one over the other. UAT-UK also runs a bursary route, and across both sittings this cycle it approved requests for 94.1% of the candidates who applied.

One last boundary. These documents answer what this cycle's two cohorts scored; they do not answer which window you should sit. Neither contains admissions outcomes, so this page derives no acceptance rate and never treats a percentile as a pass mark. The annual report says plainly that UAT-UK does not dictate how each institution uses test scores and that this may vary from university to university, so any claim about how many marks are enough is not coming from these documents. The scale is re-anchored every cycle, so this year's figures cannot be set against 2024/25. For dates and booking, see ESAT registration and test dates.
FAQ

Frequently asked questions

Is there an easier sitting?
No, nothing in the data supports that. Both events were transformed with the same scaling constants, and the report states this was done so scaled scores would be comparable across events; the scaled score has form difficulty removed already. The report's own example: a candidate scoring 6.5 in Chemistry, in either January or October, has higher ability than one scoring 4.2. So a gap like 4.73 against 4.19 on Mathematics 1 is a statement about two groups of candidates, not about two papers.
Is a 5.4 in January the same thing as a 7.0 in October?
No. On the official reading a 7.0 represents higher ability than a 5.4, and that holds across the two events. What is the same is the rank position: the 90th percentile on Mathematics 1 is 7.0 in October and 5.4 in January, so both sit at the edge of the top tenth of their own cohort. Equal rank is not equal ability, and what UAT-UK reports for each module is the 1.0 to 9.0 scaled score.
If October goes badly, can I sit again in January?
No. The annual report states that a test may be taken only once per admissions cycle and that resits were not permitted in the January sitting. Both windows were open to almost everyone, with a single exception: applicants to a Cambridge course with an October UCAS deadline had to sit in October, and the technical report names that rule as the reason October carried the higher volume. The choice is therefore made once, so weigh your target courses and your own readiness before you book. On the distributions themselves, January is the tighter of the two: standard deviations of 1.59 against 1.03 on Mathematics 1, 1.66 against 1.33 on Mathematics 2 and 1.70 against 1.51 on Physics.
I need Biology or Chemistry. Is there any January benchmark?
For Biology, no. Only 16 candidates took it in January and the report marks the mean, SD and every percentile on that row as NA. For Chemistry there is something, but it is thin: 332 candidates, mean 3.63, median 3.5, 90th percentile 5.7 and a maximum of 8.9, the only printed maximum in Table 5 that falls short of 9.0. The report also notes in its reliability section that the January Chemistry forms had fewer than 100 candidates each, so those reliability values may not be representative. October, by contrast, had 2,814 in Chemistry and 1,531 in Biology.
Can I compare this year's scores with last year's?
No. The report states that the scaling is re-fixed each cycle to that cycle's October ability distribution, so this year's scaled scores cannot be set directly against 2024/25. What does travel across cycles is the shape of the entry: last cycle 77% of 11,919 Mathematics 1 candidates sat in October and 23% in January, against 71% and 29% of 14,193 this cycle, with January deliveries up 47% year on year.
SOURCES

Official sources

MORE GUIDES

Other ESAT preparation guides

FrontierVUE is an independent practice platform. It is not affiliated with or endorsed by UAT-UK, Pearson, OCR, the University of Cambridge, Imperial College London, or any official admissions-test owner.

FrontierVUE 是独立的备考练习平台，与 UAT-UK、Pearson、OCR、剑桥大学、 帝国理工学院或任何官方入学考试主办方均无隶属或背书关系。

Put it into practice

60 ESAT papers on FrontierVUE, including 29 Frontier Original mocks built to the current format, every question with a worked solution. The first 2 papers are free.

Practise ESAT papersSee pricing