EPSO Abstract Reasoning: What the Official Sources Actually Say

If you have applied for EPSO/AD/427/26 and want to know what the abstract reasoning test actually is, the answer turns out to be unusually precise. The Notice of Competition fixes the length, the timing and the scoring. EPSO's own published sample material fixes the item format. Between the two there is very little left to guess about.

That matters, because a good deal of the abstract reasoning advice aimed at EPSO candidates describes a test that EPSO does not set. This article works only from the official notice and from EPSO's published samples, and it is explicit about where something has not been published at all.

PassEPSO is an independent preparation platform. It is not affiliated with EPSO or with any EU institution.

The short version

• Ten questions, ten minutes, scored out of 10.
• Taken in your Language 1.
• Held and invigilated remotely, at an appointment you book yourself.
• Pass/fail only. It contributes nothing to your ranking.
• Every official sample item is the same thing: five frames, choose the sixth, five options labelled A to E, in black and white.

The rest of this page is the evidence for each of those lines, and what follows from them.

What the Notice of Competition specifies

Abstract reasoning is one of six test areas in EPSO/AD/427/26, alongside verbal reasoning, numerical reasoning, EU Knowledge, Digital Literacy and the EUFTE written test. The notice sets out the three reasoning tests as follows.

Test	Questions	Time	Language
Verbal reasoning	20	35 minutes	Language 1
Numerical reasoning	10	20 minutes	Language 1
Abstract reasoning	10	10 minutes	Language 1

Abstract reasoning is scored from 0 to 10. All three tests are taken in Language 1, which is worth registering properly: abstract reasoning contains no language content at all, but the interface, the instructions and the booking correspondence around it are in your first language.

The tests are held remotely and invigilated remotely. Candidates book an individual appointment inside a limited window rather than sitting at a fixed national hour. EPSO communicates the detailed testing arrangements when it issues the invitations. No date for the tests has been officially published, and anyone telling you a month is guessing.

Takeaway

Ten questions in ten minutes is one minute per item, with nothing left over. That single number should shape your practice more than any other fact on this page.

The scoring fact most candidates miss

Abstract reasoning is a hurdle, not a ranking component.

The pass mark is not applied to abstract reasoning on its own. It is a combined 10/20 across the numerical and abstract tests together. Ten marks are available in each, twenty in total, and you need half of that total.

More importantly, the scoring table in the notice (Table 6) excludes both the numerical and the abstract test from the preliminary combined score and from the final combined score. Neither one earns you a single ranking point. Verbal reasoning is treated completely differently: it carries 35% of the final combined score.

So the abstract test can eliminate you, and it can never help you. There is no such thing as a competitively strong abstract score in this competition — only a safe one.

Takeaway

Two consequences follow. First, practise abstract together with numerical, because the pass mark is shared between them and a weak side has to be carried by the strong one. Second, once you are reliably clear of the combined threshold under real timing, stop. The marginal hour is worth far more spent on verbal reasoning, which is actually weighted.

What every official sample item looks like

EPSO publishes sample abstract reasoning items on its EU Careers site, both as an interactive test and as downloadable PDF sample sets for the AD and AST profiles. We went through all twenty published abstract items across those sets. They are strikingly uniform.

One format, on every single item

Every official item shows a series of five frames and asks you to choose the sixth. The stem is identical, word for word, on all of them:

"From the options available, please choose which diagram would come next in the series."

There is no second question type. Not one item deviates.

Five options, labelled A to E

Every published item offers exactly five answer options, lettered A to E. Not four. Not six. This is a small detail with a real consequence: a blind guess is worth one in five, and eliminating even one option changes that materially.

Black and white line art

The figures are pure monochrome line drawings. No colour has ever appeared in official EPSO abstract material. Where a rule involves fill, it is expressed through black, white and hatching, not through hue. If you are practising on colourful items, you are training your eye on a distinction the real test does not make.

Three item types that do not appear

This is where the gap between the official material and the wider aptitude-testing world is widest. In the general psychometric market, three formats are close to standard:

• Matrix items — a 3×3 grid with the bottom-right cell missing.
• Odd-one-out items — five figures, four of which share a property.
• Analogy items — A is to B as C is to what?

None of the three appears anywhere in official EPSO abstract material. Not in the interactive sample test, not in the AD sample set, not in the AST sample set. The published corpus is series items and nothing else.

Plenty of general-purpose preparation material assumes otherwise, because those formats are what the broader industry produces. Practising them is not harmful in itself — the underlying rules overlap heavily — but it is slower, and it trains a scanning habit that does not match the one thing you will actually be asked to do sixty seconds at a time.

Takeaway

Judge any practice set by a single question: is every item a five-frame series with five options? If not, it is modelling a different exam.

The format has not changed in over fifteen years

EPSO's 2009 official sample PDFs, produced for EPSO by an external test publisher, are still online. Put them beside the current material and the continuity is almost eerie: the same word-for-word stem, the same five frames, the same A-to-E options, the same monochrome style.

The item format has been stable for more than fifteen years. That is genuinely useful information. It means older official samples are still representative practice, and it means the format is unlikely to be quietly reinvented between now and your test appointment.

What did change: 2019 versus 2026

The format held steady. The weight did not.

	AD5 2019	AD5 2026
Questions	20	10
Time	20 minutes	10 minutes
Role in scoring	Counted towards ranking	Pass/fail gate only

In the 2019 AD5 competition, abstract reasoning was twenty questions in twenty minutes and it counted towards your ranking. It is now half the length and worth zero ranking points.

Preparation advice written for the 2019 competition, or copied from it, will therefore give you a study weighting that is wrong for 2026 — usually too heavy. If a study plan tells you to grind abstract reasoning for its scoring value, it is describing a competition that ended seven years ago.

EPSO publishes no explanations

This is the honest frustration of the official material. EPSO releases sample items, but no worked solutions of any kind. The interactive version tells you whether you were right or wrong. The PDF "solution" slide is the same question with a box drawn around the correct option. There is no stated rule, no reasoning, no commentary.

Which means the official samples are excellent for calibrating the format and terrible for learning from your mistakes. If you get one wrong, EPSO will not tell you what the pattern was. You have to reconstruct it yourself, or work with material that explains its reasoning.

Takeaway

Use the official samples to confirm what the test looks like and how ten minutes feels. Use explained practice to actually improve. They are two different jobs.

What EPSO says the test measures

In EPSO's own words, abstract reasoning "assesses your ability to identify logical patterns and relationships between concepts, without relying on language, numbers, or spatial orientation."

The final clause is the interesting one. This is not a spatial-rotation test in the mechanical-aptitude sense, and it is not asking you to fold shapes in your head. It is asking you to spot a rule and apply it. The figures are a delivery mechanism for the rule, nothing more.

The same test appears in the AD7 competitions

Abstract reasoning appears with identical parameters — ten questions, ten minutes — in EPSO/AD/428/26 (AD7 Audit) and EPSO/AD/429/26 (AD7 ICT). If you are preparing for one of those competitions, everything on this page applies to you as well.

What to do with all this

Practically, the whole page reduces to four instructions.

1. Practise the five-frame series format, and only that format, under a one-minute clock.
2. Practise it in mixed sets with numerical reasoning, because that is how the pass mark is calculated.
3. Aim for a comfortable margin above the combined threshold, not for a perfect score that earns nothing.
4. Learn to name the rule families, so that a missed item becomes a diagnosis rather than a shrug.

That last point is the one that takes real work, and it is the subject of the companion article: the pattern types and rule families worth learning for EPSO abstract reasoning.

Official sources

• Notice of Competition EPSO/AD/427/26, Official Journal C/2026/711, 5 February 2026
• EU Careers — Abstract reasoning, including the interactive sample test
• Notice of Competition EPSO/AD/428/26 (AD7 Audit) and EPSO/AD/429/26 (AD7 ICT, Official Journal C/2026/2425)
• Notice of Competition for AD5 2019, Official Journal C 177 A, 23 May 2019

Always treat the EUR-Lex notice and the EU Careers pages as the binding source for anything EPSO asks you to do.