Employers often use a vendor's own cognitive test. Here is what public sources say about 12 of them, and how to practise the question types.
Most vendor cognitive tests are short, timed and mixed. They usually combine number, word and pattern questions, and many are designed so that most people do not finish. The specific names differ, but the skills are shared: working accurately under time pressure, spotting rules in sequences and shapes, and reading short texts carefully.
This page covers 12 tests one by one, with the format as far as public sources describe it, a worked example, common mistakes and a practice plan. Where sources disagree, we say so instead of guessing. Confirm the format in your own invitation.
| Test | Typically used for | Main question types |
|---|---|---|
| RCAT | Graduate and general hiring, widely reported in Australia and New Zealand | Verbal, numerical, abstract |
| PI Cognitive | Hiring and talent decisions across many industries | Verbal, numerical, abstract |
| PXT Select | Role-fit screening with benchmarks | Verbal, maths; plus behaviour and interests |
| Prevue | Hiring across many roles, with a timed ability test | Numbers, words, shapes |
| MMAT | Pre-employment screening of mental agility | Maths, vocabulary, logic |
| AssessFirst BRAIN | Estimating reasoning and learning ability for hiring | Numerical, verbal, abstract |
| Bryq | Pre-hire screening, usually combined with a personality section | Numerical, verbal, logical, attention to detail |
| Omnia | Measuring general mental ability in hiring | Verbal, numerical |
| Harver | High-volume hiring, with employer-chosen subtests | Logical, numerical, verbal, spatial, perceptual |
| Caliper | Pre-employment assessment that is mostly personality | Abstract reasoning (number and shape series) |
| Metis ACE | Measuring reasoning ability with an adaptive test | Numerical, spatial, verbal |
| BairesDev | Screening for technology roles | Maths word problems, estimation, situational |
Descriptions as of 13 October 2026, from the sources named in each section. Formats vary by employer and version, so check your invitation.
The Revelian Cognitive Ability Test, known as RCAT and formerly Onetest, is a timed test of verbal, numerical and abstract reasoning. Test-prep sources describe it as short and fast, with difficulty rising as you go.
Test-prep sites describe 51 questions in about 20 minutes, split roughly evenly across verbal, numerical and abstract items. They report one point per correct answer and no penalty for wrong answers, and say most candidates do not finish. One site lists a 25-question, 10-minute sample, which is a practice version and not the live test.
Reported verbal items include synonyms and antonyms, word swaps, analogies and word logic. Numerical items are described as calculations done on paper, some as fill-in answers instead of multiple choice. Abstract items show five or six pictures and ask you to find the two that do not follow the pattern.
Speed and accuracy across three skills: vocabulary and word relationships, mental arithmetic and number logic, and spotting rules in visual series. Revelian describes the test as one that does not rely on specific knowledge, so the gains come from familiarity and pacing.
Pens cost 3 for £2.40. How much do 10 pens cost at the same price?
Answer: £8.00
Three pens cost £2.40, so one pen costs £0.80. Ten pens cost 10 × £0.80 = £8.00.
Upload a screenshot of a practice abstract or numerical question and compare your own working with the explanation. Answer-only mode suits timed rehearsal; Tutor mode suits new patterns.
Abstract reasoning solver Numerical reasoning solver Revelian abstract reasoning page
Test-prep sites describe 51 questions in about 20 minutes, but they are not Revelian's own documentation and some sources differ. Check your invitation.
Test-prep sources report one point per correct answer and no penalty for wrong answers. Confirm this in the instructions shown at the start of your test.
Source basis, as of 13 October 2026: test-prep sites and candidate guides (no Revelian page was read). Sources disagree on some details, so treat the numbers as indicative.
The Predictive Index Cognitive Assessment is a short timed test of reasoning. It is separate from the PI Behavioral Assessment, which is an untimed adjective checklist, so check which one your employer has set.
Test-prep sources describe a timed test of about 50 questions in around 12 minutes, covering verbal, numerical and abstract reasoning, which works out at roughly 14 seconds a question. They say very few candidates answer all of them. The split between sections differs from site to site, and we could not confirm it against Predictive Index's own documentation.
Abstract items are described as finding patterns and relationships in shapes or sequences. Numerical items are described as number series and short problems; verbal items as word relationships and short reasoning.
Rapid pattern recognition and quick, accurate reasoning with limited time. With about 14 seconds a question, the skill is deciding fast whether to answer now or move on.
Find the next number: 4, 7, 12, 19, 28, ?
Answer: 39
The gaps are 3, 5, 7, 9, so they grow by 2 each time. The next gap is 11, and 28 + 11 = 39.
Use ReasonEra to check your number and shape patterns in timed practice, then read the explanation for any rule you missed. Repeating short sets builds the speed this test rewards.
Numerical reasoning solver Inductive reasoning solver Pacing lessons from the CCAT
Test-prep sources describe about 12 minutes and 50 questions. These are third-party descriptions, so confirm the time in your invitation.
No. It is described as an untimed adjective checklist about work style, so it has no right or wrong answers in the way a reasoning test does.
Source basis, as of 13 October 2026: test-prep sites (no Predictive Index page was read). Section counts and average scores quoted online could not be verified.
PXT Select is a role-fit assessment that combines a thinking-style (cognitive) section with behavioural traits and interests. This page covers only the cognitive part, which is the section with right and wrong answers.
Vendor and directory descriptions say PXT Select measures cognitive ability, behavioural traits and interests, and compares candidates with a role benchmark. The behavioural part asks you to agree or disagree with statements; we do not cover it here.
The cognitive part is described as verbal content, such as synonyms, antonyms and analogies, plus maths. Sources conflict on size and timing: one prep site reports 77 untimed cognitive items inside a roughly 60-minute assessment, and another reports about 50 questions in around 20 minutes. Treat both as unconfirmed.
Vocabulary and word relationships, basic arithmetic and number reasoning. Because the result is read against a role benchmark, the score is judged for that job, not as a universal number.
Which word is closest in meaning to 'scarce'?
Answer: rare
Scarce means in short supply, which is closest to rare. Plentiful is the opposite.
Practise the reasoning and number questions with ReasonEra, and use the explanations to see why a wrong option fails. This builds the habit of checking the relationship, not the first familiar word.
Verbal reasoning solver Numerical reasoning solver Verbal reasoning tips
No. They are different assessments from different vendors, although some prep pages confuse them.
Sources conflict. One prep site says untimed, another describes about 20 minutes. Check your invitation.
Source basis, as of 13 October 2026: vendor listings and test-prep sites. Item counts and timing conflict between sources and are unconfirmed.
Prevue's ability test is a timed test covering three areas: working with numbers, working with words and working with shapes. Prevue's help centre says its ability tests are timed and that wrong answers are not penalised.
Prevue's help centre says the instructions for each test show how many minutes you have to complete as many questions as you can, and that you are not penalised for wrong answers. Third-party prep sites describe about 94 questions in 18 to 20 minutes, in four sections that mix free-response and multiple-choice items; Prevue's own pages did not give a total, so that figure is unresolved.
The numbers section is described as arithmetic, number sequences and numerical analogies. The words section is described as hidden words and letter sequences. The shapes section is described as completing patterns of shapes.
Speed on short, varied items: number relationships, word manipulation and pattern completion. With so many questions in so little time, the skill is moving fast and not stalling.
3 is to 27 as 4 is to ?
Answer: 64
27 is 3 cubed (3 × 3 × 3). Applying the same rule, 4 cubed is 64.
Use ReasonEra on number and shape practice questions to check the rule you spotted. Short, repeated sets help with the speed this test depends on.
Numerical reasoning solver Logical reasoning solver Numerical reasoning tips
Prevue's help centre says you are not penalised for wrong answers on its ability tests. Check your own instructions.
Prep sites say 18 to 20 minutes for about 94 questions, but Prevue's own pages did not confirm the total. The instructions show the time for your version.
Source basis, as of 13 October 2026: Prevue help centre. Question counts and section timings come from third-party prep sites and are unresolved.
The McQuaig Mental Agility Test, or MMAT, is a short, timed test of mathematical ability, vocabulary and logical reasoning. It is one of McQuaig's tools, separate from its personality-style word survey.
Prep and test-information sites describe 50 multiple-choice questions in 15 minutes, in a mixed order, usually taken online with a timer on screen. They say calculators and phones are not allowed and that you may use blank paper for calculations. Most people are not expected to finish.
Results are described as percentiles against a comparison group. We found no published pass mark, and employers compare results with their own role needs. Average scores quoted on prep sites could not be verified.
Quick mental arithmetic, vocabulary and simple logic, all under tight time. About 18 seconds a question is the average budget if the 50-question, 15-minute figures are right.
If 5 machines make 5 parts in 5 minutes, how long would 100 machines take to make 100 parts?
Answer: 5 minutes
Each machine makes one part in 5 minutes. So 100 machines make 100 parts in the same 5 minutes.
Practise logic and number questions in ReasonEra and read the explanation for the ones that trick you. It is useful for learning the trap in puzzles like the machines question.
Logical reasoning solver Numerical reasoning solver CCAT pacing tips (similar format)
Public descriptions say 50 multiple-choice questions in 15 minutes. Confirm in your own instructions.
We found no published pass mark. Results are shown as percentiles and employers judge them against their own roles.
Source basis, as of 13 October 2026: McQuaig blog page on the MMAT and third-party summaries. The 50-question, 15-minute format and any average scores come from third-party summaries.
BRAIN is AssessFirst's reasoning test. AssessFirst says it evaluates reasoning and learning ability and gives an overall cognitive potential score from 1 to 10, with a breakdown of reasoning abilities.
AssessFirst's own material describes a short assessment, and its help centre says assessments typically take 5 to 10 minutes. An unofficial prep site describes an older 21-minute version with four mini tests: numerical (tables and graphs), verbal (short passages with true, false or cannot-say statements, plus analogies) and abstract (shape series), with difficulty rising in each.
A French guide says the earlier test took about 20 minutes and was redesigned to be shorter and mobile-friendly, so the current format may differ from older descriptions.
Reading data tables and charts, careful reading of short passages, and finding rules in shape series, all in a compressed time.
Passage: 'The library extended its weekend hours in March after a survey found most visitors came on Saturdays.' Statement: 'The library is now open on Sundays.' True, false or cannot say?
Answer: Cannot say
The passage says weekend hours were extended but does not say Sunday opening was included, so the statement cannot be confirmed.
Practise passage and pattern questions in ReasonEra and use the explanation to see which sentence supports the answer. It trains the judgement this test asks for.
True, false, cannot say guide Verbal reasoning solver Abstract reasoning solver
AssessFirst's help centre says assessments typically take 5 to 10 minutes. Older prep descriptions of about 21 minutes appear to refer to an earlier version.
AssessFirst describes an overall cognitive potential score from 1 to 10, reflecting reasoning and learning ability, with a breakdown by reasoning area.
Source basis, as of 13 October 2026: AssessFirst help centre. Section details come from an unofficial prep site and may reflect an older version.
Bryq's cognitive assessment measures four areas: numerical reasoning, verbal reasoning, logical reasoning and attention to detail. Bryq says the items are scenario-based, timed, and not tied to any job's hard skills.
Bryq's help centre describes a Cognitive Only version of 20 questions with a nominal duration of 7 minutes. Its candidate page says items are designed to be simple but timed, and that if you do not answer in time you simply move on to the next question. The standalone cognitive and extended versions are noted as deprecated, available only to existing customers, so most candidates now take a combined cognitive and personality assessment.
Numerical items draw conclusions from numerical data, verbal items evaluate written information, logical items solve problems from observed patterns, and attention-to-detail items ask you to spot details in a lot of information. Results are compared with role-specific benchmarks, and Bryq advises reading the skill profile as well as the total.
Clear thinking with everyday data and text, fast pattern spotting, and careful checking, in a scenario format rather than abstract puzzles alone.
A team resolved 42 tickets on Monday, 35 on Tuesday and 49 on Wednesday. What was the average per day?
Answer: 42
The total is 42 + 35 + 49 = 126. Divided by 3 days, the average is 42.
Use ReasonEra for pattern and reasoning practice, then apply the same checking habits to data and text scenarios. The explanations help you see exactly where a quick read went wrong.
Logical reasoning solver Data interpretation tips Error-checking tests
Bryq describes four areas: numerical reasoning, verbal reasoning, logical reasoning and attention to detail.
Bryq's help centre gives a nominal 7 minutes and 20 questions for the Cognitive Only version, but most candidates now take a combined assessment. Check your invitation.
Source basis, as of 13 October 2026: Bryq help centre: cognitive assessment overview. These are Bryq's own descriptions; we found no independent validation of the format.
The Omnia Cognitive Assessment measures general mental ability, which Omnia describes as the ability to reason, solve problems, think abstractly and learn quickly. Omnia says it takes about 15 minutes.
Omnia's materials describe a timed test, online or on paper and proctored, with two subtests: verbal, which measures knowledge of words and verbal concepts, and numerical, which measures maths, reasoning and problem solving. A third-party prep site reports 57 multiple-choice questions in total, which we could not confirm on Omnia's pages.
Results are given as percentiles, meaning the share of the comparison group with fewer correct answers, and employers receive a one-page report straight away.
Vocabulary and verbal relationships, plus numerical reasoning that includes percentages and everyday maths, under time pressure.
A price rises from 80 to 92. What is the percentage increase?
Answer: 15%
The increase is 92 − 80 = 12. As a share of the original price, 12 ÷ 80 = 0.15, which is 15%.
Practise numerical and verbal reasoning questions in ReasonEra and read the explanation for each wrong answer. It is a way to find which of the two subtests needs more work.
Numerical reasoning solver Verbal reasoning solver What is a good test score?
Omnia's materials describe about 15 minutes. Your invitation will give the exact time.
Omnia describes it as an ability to reason, plan, solve problems, think abstractly, understand complex ideas and learn from experience.
Source basis, as of 13 October 2026: Omnia Group: the Omnia cognitive assessment. The 57-question figure comes from a third-party prep site and is unconfirmed.
Harver's cognitive assessment is modular, so employers choose which subtests to include for each role. Harver lists eight subtests and says each takes roughly 5 to 15 minutes.
Harver describes subtests such as logical reasoning (pattern recognition in figures), numerical reasoning (finding the system in number sequences), verbal reasoning, abstract reasoning, perceptual speed and accuracy, spatial ability, number patterns and reading comprehension. Candidates take them online.
Harver says scores are shown on a 0 to 100 norm-referenced scale, where 50% is the average and most people score between 30% and 70%. A third-party prep site shows logical items as shape analogies where you apply a transformation, such as rotation or duplication; treat that as illustrative.
Depends on the subtests chosen: pattern recognition, number-sequence logic, accuracy at speed and spatial thinking. Because the mix is employer-specific, flexibility matters more than depth in one area.
Find the next number: 2, 3, 5, 8, 12, ?
Answer: 17
The gaps are 1, 2, 3, 4, so the next gap is 5, and 12 + 5 = 17.
Use ReasonEra for the logical and numerical pattern types and read the explanation to find the transformation or rule. It is a quick way to build the rule-naming habit these subtests reward.
Inductive reasoning solver Abstract reasoning solver Spatial reasoning solver
It depends on the employer and role. Harver says its cognitive tests are modular, with eight subtests available.
Harver says results are on a 0 to 100 scale compared with a norm group, where 50% is the average and most people score between 30% and 70%.
Source basis, as of 13 October 2026: Harver: cognitive ability testing. Sample-item descriptions come from a third-party prep site and are illustrative.
The Caliper Profile is mainly a personality-style questionnaire, so this section covers only its reasoning content. Talogy describes it as using personality and cognitive-ability items to measure talent effectiveness.
Sources agree that the reasoning part is abstract reasoning and that it is small compared with the personality part. Some describe a short set of number series and shape patterns with objectively right answers. Others describe a larger timed section that includes verbal items such as synonyms and antonyms.
Timing is also inconsistent: one source says the test is untimed and takes most people 60 to 75 minutes overall, while others call the cognitive section timed. Talogy's own documentation would settle this, so confirm with your employer.
Finding the rule in a number series or shape sequence, including alternating patterns that need two rules at once.
Find the next number: 2, 5, 4, 7, 6, 9, ?
Answer: 8
Two interleaved series: 2, 4, 6, 8 in odd positions and 5, 7, 9 in even positions. The next term continues the first series: 8.
ReasonEra is for the reasoning items only. Practise number and shape series there, and use Tutor mode to see how an alternating rule is broken down.
Inductive reasoning solver Abstract reasoning solver Inductive reasoning tips
It is mainly a personality-style questionnaire. Sources describe a smaller abstract-reasoning section with right and wrong answers.
Sources conflict. Check your own invitation or ask the employer.
Source basis, as of 13 October 2026: test-prep sites and Talogy's own description of the profile. Whether the reasoning section is timed, and whether it includes verbal items, is inconsistent across sources.
ACE is Master International's cognitive ability test on its Metis assessment platform. Master describes it as a test of reasoning ability that predicts performance and future success. Not to be confused with the Addenbrooke's Cognitive Examination, a clinical screening tool.
Descriptions say ACE covers three domains: numeric, spatial and verbal. It is described as computer-adaptive, so the difficulty of the next question depends on how you answered the previous ones. One prep site reports 20 to 28 questions depending on the version, with a time limit of 60 to 75 minutes, and says answers may be multiple choice or free response.
Master's pages describe the test as measuring logical, analytical reasoning and the ability to see patterns and complex relationships. Question counts and time limits vary by version, and the publisher has moved to a newer platform name, so check what your employer uses.
Steady accuracy across numbers, shapes and words. In an adaptive test, getting answers right raises the difficulty, so feeling that questions are getting harder is normal.
A cube is painted on all faces and then cut into 27 equal small cubes (3 × 3 × 3). How many small cubes have exactly two painted faces?
Answer: 12
Cubes with exactly two painted faces sit on the edges but not at the corners. A cube has 12 edges, and each has one such middle cube, so 12.
Practise spatial and numeric puzzles in ReasonEra, then use the explanation to understand the counting or visualising step. It helps with problems like the cube question that reward a method.
Spatial reasoning solver Numerical reasoning solver Logical reasoning solver
Descriptions say it is computer-adaptive, so question difficulty adjusts to your answers.
Master International's descriptions point to three domains: numeric, spatial and verbal reasoning.
Source basis, as of 13 October 2026: Master International: ACE overview. Question counts and time limits come from a prep site and vary by version.
BairesDev does not publish a format that we found, so this section rests on candidate reports. They describe timed, multiple-choice problem-solving tests as part of a wider screening process.
Candidate reports describe several linked tests: practical skills, problem solving, advanced problem solving, situational questions and a Predictive Index assessment. Problem-solving tests are described as intermediate to advanced, with maths word problems of a college-entrance style and educated estimates on general topics, one reporting an 8-question set.
Candidates describe strict timing and, in some cases, a live problem-solving exercise on a screening call. One reports a six-month wait before retaking. These are user-generated accounts, so details may differ for you; ask the recruiter for the current format.
Structured reasoning on word problems, estimation (Fermi-style), and staying accurate under strict timing.
A team of 6 developers closes 120 tickets in 5 days. At the same rate per developer, how many tickets would 9 developers close in 8 days?
Answer: 288
Each developer closes 120 ÷ (6 × 5) = 4 tickets a day. Then 9 developers × 8 days × 4 tickets = 288.
Practise word-problem and logic questions in ReasonEra and use the explanation to see the per-unit step. It is a way to learn the method, which carries over to unfamiliar problems.
Logical reasoning solver Numerical reasoning solver How to pass a psychometric test
Candidate reports describe timed, multiple-choice problem-solving tests that include maths word problems and estimation. BairesDev did not publish a format that we found.
One candidate reported a six-month wait. This is user-reported, so confirm with the recruiter.
Source basis, as of 13 October 2026: candidate interview reports (not official BairesDev material). All details are user-reported and may be out of date.
ReasonEra is a specialised AI for reasoning-test questions. A tool made for puzzle patterns can be quicker and more focused than a general-purpose chatbot. See our own published benchmark for the method and results.
We maintain the model and keep developing it, adding new puzzle types and question formats as they appear.
Many candidates now use AI to study. ReasonEra gives you an AI tool for reasoning tests, with explanations that teach the pattern and not only the answer.
It is a reasoning test built and sold by an assessment company and used by many employers, as opposed to a test an employer writes itself. Most are short and timed, and most mix number, word and pattern items.
No, but they overlap a lot. Most test verbal, numerical and abstract reasoning under time pressure. Format, length, scoring and benchmarks differ, so confirm yours in the invitation.
Practise the shared skills: number series, percentages and ratios, word relationships, passage-based true or false questions and pattern rules. Add timed mixed sets, because most of these tests are built so that many people do not finish.
Treat them as indicative. Each section names its basis, and several rest on test-prep sites or candidate reports, which can be out of date or disagree. Your invitation is the authority.
ReasonEra is built for reasoning-test questions. You can practise the underlying question types, such as pattern, logic and number questions, and see the explanation. It is for practice and learning, not for use during a live assessment.
Try a few practice questions in ReasonEra with the free trials, and see the logic behind each answer.
Billing is handled by Paddle, and refunds follow the Refund & Cancellation Policy.
Notes. Facts on this page come from public pages and test-prep reports, and each section names its basis. Test formats change, so always confirm details in your own invitation. ReasonEra has no affiliation with the organisations named on this page, and the test names belong to their owners. ReasonEra is for practice and learning only.