Test Partnership Practice Tests
Insights, Concepts, Perceptions, TPAQ and the MindmetriQ games, built to the real item counts and the real clocks. Every test in this range is computer adaptive, and that changes what your practice score means.
Free Tests
Free to try with a free account - unlock this provider for unlimited practice.
Insights Numerical Reasoning
Five question sets, each a chart or table with three questions and one shared three-minute budget. Ten near-miss options per question and a calculator expected, matching the Test Partnership Insights numerical test.
MindmetriQ Net the Numbers
Move a net across a grid of numbers to capture the highest possible total. The grid contains negatives, so the tile that looks biggest is often the wrong place to land.
Premium Tests
Insights Verbal Reasoning
Five passages, four statements each, judged on the five-point Definitely True to Definitely False scale. Three minutes covers reading the passage and all four statements.
Insights Inductive Reasoning
Fifteen figure sequences with the next frame missing, sixty seconds each and no borrowing time from the next item. Adaptive, with difficulty tuned to the 70% success rate Test Partnership targets.
Concepts Critical Thinking
The high-stakes test in the range, and four item types in one sitting: constraint deduction, conclusion evaluation, argument strength on a six-point scale, and inference. Each block carries its own clock.
Perceptions Situational Judgement
Untimed workplace scenarios with four actions each, rated on a six-point effectiveness scale that has no neutral middle. Credit spans two adjacent labels, so calibration matters as much as ranking.
TPAQ Personality Profile
Test Partnership’s personality questionnaire, practised across the nine global factors of its Big5+ framework including Integrity. Untimed, seven-point agreement scale, and no score to beat.
MindmetriQ Link Swipe
Classify each pair of words as synonyms, antonyms or neither, against a short per-item clock. Swipe on a phone or tap on any device.
MindmetriQ Pipe Puzzle
Rotate pipe segments to join two endpoints around solid obstacles, with a limited supply of pipes. Planning the route before you touch anything is the whole skill.
Does your invitation say Concepts or Insights? It is the same test
Verified September 2026The rebrand is half finished.On one page, right now, the body copy says "Insights Critical Thinking Test", the factsheet it links to is titled "Concepts Critical Reasoning", and a customer testimonial on the same page says "Concepts Critical Thinking". Whichever name your invitation uses, the test is the one we build here.
Customers drop the plural too. Insight and Insights are the same suite.
If the link goes to tptests.com, you are in the right place. That is Test Partnership's own candidate platform, and it carries the instruction screens every figure on this page comes from.
Mechanical reasoning is now branded "Insights Mechanical" on the website but missing from the Insights factsheet, so treat it as real and recently renamed. We do not build it yet. TPAQ appears only as an acronym in every piece of Test Partnership material we could find, so we do not state what it stands for.
These tests are built so you get about 70% right, whoever you are
Why 70% and not 50%
A textbook adaptive test aims each question at a 50% chance of success, because that extracts the most information. Test Partnership deliberately does not. Its technical manual says it administers questions with a 70% chance of success so that "candidates will typically answer 70% of their questions correctly", accepting a small loss of precision in exchange for a test that does not leave everyone feeling they failed. It offsets the loss by adding one or two items.
So your accuracy is not the signal
A strong candidate and a weaker one can both finish feeling they got seven in ten right, because the test moved to meet each of them. What differs is the difficulty they were answering at when it settled. That is also why our scoring weights the difficulty you cleared above the count you cleared: Test Partnership does the same, and says so. "The score weighting of each question is determined by its difficulty."
And guessing is correct when stuck
There is no negative marking anywhere in this range, and Test Partnership tells candidates to use their best guess rather than let a question expire. An expiry is a guaranteed zero; a guess is not. The one thing that makes guessing a poor strategy is the option design: the numerical test offers ten options spaced a tenth of a percentage point apart.
Three different clocks, and the one most candidates get caught by
On the numerical and verbal tests, the clock belongs to the question set, not to the question. Test Partnership's instruction screen puts it plainly: "This test comprises five question sets. Each question set has a time limit of three minutes and contains three questions. Therefore in total, you will have 15 minutes to answer 15 questions." The header on its sample reads "Time remaining for these questions", plural.
That is a different skill from a per-question limit and it creates a failure mode no other provider in our library can. Spend 150 of your 180 seconds understanding the chart and the second and third questions are gone, marked unanswered, regardless of whether you could have answered them. The fix is a reading habit rather than a speed habit: read the chart or passage once, for its structure, its units and its footnotes, before you look at any of the questions. That reading serves all three.
Our simulations implement this exactly, including failing the rest of a set when its budget expires, because practising it on a per-question clock would teach the wrong pacing.
One budget per question set
Insights Numerical (3 minutes per chart and its three questions) and Insights Verbal (3 minutes per passage and its four statements). Concepts works the same way but the budget varies by item type: 120 seconds for a three-statement deduction, 90 for a two-statement conclusion, 60 for a single argument or inference.
One clock per question
Insights Inductive gives 60 seconds to each of its 15 figure sequences with nothing carried forward, and the MindmetriQ games run 15 to 30 seconds an item. Their anti-cheat page treats that short clock as a security feature in itself: there is no time to copy a question out and get an answer back.
No clock at all
The Perceptions situational judgement test and the TPAQ personality questionnaire are both untimed, and Test Partnership is emphatic about it. Rushing either gains you nothing. Our versions show no timer, because showing one would train exactly the wrong behaviour.
The formats, and what each one actually asks
Insights Numerical
15 questions in 15 minutes, as five sets of three sharing one three-minute budget each. Ten answer options per question, deliberately spaced a tenth of a percentage point apart so a guess is a poor bet. A calculator and rough paper are expected, so practising mental arithmetic is preparing for a constraint that does not exist.
Insights Verbal
Five passages of four statements, three minutes per passage. The scale is five points: Definitely True, Probably True, Insufficient Information, Probably False, Definitely False. "Definitely" means beyond reasonable doubt from the passage alone. Marketing pages disagree on the item count; we build the 20 the instruction screen states.
Insights Inductive
Fifteen figure sequences, 60 seconds each, no borrowing from the next item. Adaptation happens after every question because every question is its own stimulus. If you have not found the rule in the first twenty seconds, commit and move on: the item is gone either way when the clock stops.
Concepts Critical Thinking
The high-stakes test in the range: 24 questions and four item types in one sitting, each with its own response scale. Constraint deduction on three points, conclusion evaluation on two, argument strength on six with no neutral option, and inference on the five-point confidence scale. Twelve norm groups against the Insights suite’s three.
Perceptions Situational Judgement
Untimed. Each scenario lists four actions and you rate every action on six points from Very ineffective to Very effective, with no neutral middle. Credit spans two adjacent labels rather than one exact point, so being one step out still scores. Five role variants exist: Graduates, Managers, Sales, Administration and Customer Service.
TPAQ personality
180 adaptive items, untimed, typically 10 to 20 minutes, measuring 45 traits and nine global factors that Test Partnership calls Big5+. The "+" is Integrity: Fairness, Greed Aversion, Honesty and Modesty, which most personality inventories have no equivalent of. There is no overall score unless the employer builds a JobFitScore.
MindmetriQ, the six games
Net the Numbers and Number Racer for numerical, Link Swipe and Word Logic for verbal, Pipe Puzzle and Shape Spinner for logical. Per-item clocks, adaptive from calibrated pools, and real physical interactions: drag, rotate, swipe, hold and release. Fully mobile-supported, unlike the traditional tests.
Verification tests
A shorter, supervised re-sit of a test you have already taken remotely, usually at an interview day or assessment centre. The employer compares the two scores. This is a shipped capability, listed as a key feature of the Insights range, and it is the reason an inflated remote score becomes a problem rather than an advantage.
The longer verbal variant
A live practice screen for a test titled simply "Verbal Reasoning" states 32 questions in 24 minutes, with the identical five-point scale and identical passage-plus-four-statements structure. That is eight passages at three minutes each. If your invitation says 24 minutes rather than 15, you are sitting that one rather than the Insights test.
Tests in this range we do not simulate, and why
Test Partnership sells around twenty distinct assessments. We build nine. These are the ones you could be invited to that you will not find here, with the honest reason in each case, because a candidate who knows a stage exists and knows we do not cover it is better off than one who is handed something that looks close.
Concepts Data Analysis
A real test and a real stage: 24 questions in 24 minutes on high-level quantitative data, computer adaptive, with twelve norm groups. We do not simulate it and we are not going to guess. There is no sample test, no product page and no live instruction screen for it anywhere on Test Partnership’s site, so its item presentation, its question grouping and its clock are simply not established. Even the item count is contested: the factsheet says 24 in 24 minutes and the range page says 20 in 20. Building it by analogy with the numerical test would mean inventing the mechanics and telling you they were real. If you are invited to it, the closest honest preparation here is Insights Numerical, and you should expect the real thing to differ.
Insights Mechanical
Fully specified and genuinely buildable: 24 questions in 20 minutes, one animated scenario per item, five options, 60 seconds each, and applied physics in real settings rather than formula recall. It is deferred rather than blocked, because its content is narrow (trades, engineering and emergency services) and we cover the same construct in our mechanical reasoning test. It is the first candidate for a follow-up.
Number Racer, Word Logic and Shape Spinner
The other three MindmetriQ games. We build one per cognitive area, which is Test Partnership’s own recommended minimum battery, so that each of the three gets proper scoring, proper feedback and a proper phone layout. Word Logic is verbal analogies and Shape Spinner is a rotation puzzle where every symbol must sit next to one sharing its shape or its shading but not both. Number Racer is real-time falling objects and is the hardest of the six to make good on a small screen.
The other TPAQ variants, and the soft-skills range
TPAQ-27 Express, the TPAQ Competency Profile, the TPAQ-Leadership Profile, the Emotional Intelligence Test, Strengths Based Assessments and Role-Specific Soft Skills. Item counts are published for all of them, but no response format, no sample items and no dimension list is available for any, so there is nothing to build to. Practising the TPAQ profile here covers the same self-report format at a workable length.
The new mobile-first verbal format
Test Partnership is piloting a verbal test titled "Verbal Reasoning - Part 1 of 3: Conclusions" that works quite differently: you read the passage on its own first, then answer four conclusions one at a time on a three-point True / Cannot Say / False scale, with the passage still visible. There is no submit button at all, and the assessment moves on when the timer ends rather than when you act. The per-item timings are behind authentication and Parts 2 and 3 are not publicly described, so we have not built it. If you meet it, the thing to know is that selecting your answer is the only input there is.
MS Word and Excel tests
30 questions in 15 minutes on practical software proficiency. Out of scope as a product rather than unresearched: a half-faithful simulation of a desktop application teaches the wrong muscle memory. Practise in the real software.
How Test Partnership scores you, and the pass mark nobody publishes
A percentile, a sten and a band
Its published sample report gives a percentile, defined as "the percentage of people's scores that your score exceeds", a sten from 1 to 10 with a mean of 5.5 and a standard deviation of 2, and a verbal band such as "Well above average". There is no published pass mark for any test in the range: the cut-off is the employer's to set.
Your rank depends on who you are compared with
MindmetriQ reports the same performance against three norm groups side by side. In the published sample, one candidate's Net the Numbers result is the 90th percentile against graduates, professionals and managers, the 88th against administrative and non-graduate staff, and the 89th against apprentices and school leavers. Concepts has twelve norm groups and TPAQ has fourteen.
Number correct is not the raw score
The sample report is explicit: "The score weighting of each question is determined by its difficulty, so difficult questions are worth more than easier questions." Two candidates with the same number correct will receive different scores if they answered items of different difficulty. Any practice that reports "8 out of 12" as the headline has misunderstood the instrument, which is why ours does not.
Phone or laptop?
It depends which test, and Test Partnership is clear about both halves. For the traditional tests its instruction screens say: "We recommend that our assessments are completed on either a desktop or laptop computer. It is not recommended to complete our assessments using a mobile device." For MindmetriQ the opposite holds, and its FAQ says the games are designed from the ground up for smaller screens; the Pipe Puzzle demo tells candidates "You can use desktop, laptop or mobile".
Every test here works properly on a phone, because revising on a train is a real thing people do. But sit the real Insights, Concepts or Perceptions test on a laptop, with a calculator and paper next to you for the numerical one. A ten-option numerical question read off a phone screen under a shared three-minute budget is a harder test than the one the employer thinks they are giving you.
Frequently Asked Questions
Is my test called Concepts Critical Thinking or Insights Critical Thinking?
Both, and it is the same test. Test Partnership is rebranding its critical thinking test from Concepts to Insights, and as of September 2026 the rebrand is half finished on a single page: the body copy of the product page calls it "Our Insights Critical Thinking Test", the factsheet it links to is titled "Concepts Critical Reasoning", and a customer testimonial on that same page says "Concepts Critical Thinking". If your invitation names either, practise the same thing.
How long is the Test Partnership numerical test, and does the clock run per question?
Neither per question nor once for the whole test. It is five question sets of three questions, and each SET has a three-minute limit, giving 15 questions in 15 minutes. Test Partnership’s own instruction screen says so directly, and the header on its sample reads "Time remaining for these questions", plural. Time you spend on the first question of a set is taken from the third.
How many questions is the Insights verbal test?
Three first-party figures exist and they disagree. The range page says 15 questions in 15 minutes, the product page says 15 questions in 20 minutes, and the live instruction screen on the delivery platform says 20 questions with a maximum of 15 minutes. We build the instruction screen’s figure, because it is what a candidate actually reads on the day and because it reconciles with the sample structure: five passages of four statements at 180 seconds each is exactly 15 minutes.
Can you use a calculator on a Test Partnership test?
On the Insights numerical test, yes, and you are expected to. The instruction screen says "You will require a calculator and some rough paper for this test. Please ensure that you have these before starting." On the MindmetriQ games, no: the technical manual states that they require basic calculations in quick succession without a calculator, and that only general addition or subtraction is needed.
Is there negative marking on a Test Partnership test?
No source anywhere, first party or third party, mentions negative marking, deducted marks or a guessing penalty in this range. The positive evidence is the instruction a candidate is given before the verbal test: "Please try to attempt every question before the end of the time limit. If you cannot answer a question in time, use your best guess." A publisher that penalised wrong answers would not say that. So always guess rather than letting a question expire.
What does "adaptive" mean here, and does the test still have a fixed number of questions?
Yes, it does. Every adaptive test in this range publishes a fixed item count (15, 20, 24, 12, 10) and a fixed total time. What varies per candidate is which items you see, not how many. Test Partnership additionally aims each item at a 70% chance of success rather than the 50% a textbook adaptive test would, so most candidates finish having answered about seven in ten correctly whatever their ability. Your raw accuracy is therefore a poor guide to how you did; the difficulty you settled at is the signal.
What is a good Test Partnership score?
There is no published pass mark for any test in the range, because the cut-off is the employer’s to set. Results come back as a percentile against a norm group, a sten score from 1 to 10 with a mean of 5.5, and a verbal band such as "Above average". MindmetriQ reports the same performance against three different norm groups side by side, so the same result can be the 90th percentile against graduates and the 88th against administrative staff.
Is the Perceptions situational judgement test timed?
No, and this is emphatic across three first-party sources: the website badge reads "Untimed", the factsheet lists "No time limits" as a key feature, and the sample test sets a per-scenario limit of 30,000 seconds, which is how it expresses "no meaningful limit". Typical completion is around 30 minutes for 50 to 70 rated actions. Taking your time costs you nothing there.
Do I sit all six MindmetriQ games?
That is the employer’s choice and Test Partnership’s own FAQ asks the question. Its guidance is one game per cognitive area for quick screening, about 12 minutes; four or five games for targeted selection, about 18 minutes; or all six plus personality for a full profile, about 35 minutes. We build one per cognitive area, which is its recommended minimum battery.
Companies That Use Test Partnership
Test Partnership is a London assessment publisher whose customers cluster in law, professional services and early careers. Several name the specific tests they use in Test Partnership’s own published testimonials, which is unusually concrete evidence and is why those entries say which test rather than only which vendor.
Eversheds Sutherland
LegalNamed in an attributed testimonial from a Senior Talent Manager. Test Partnership runs a dedicated legal sector page and law firms are its strongest customer cluster.
View Eversheds Sutherland guide →DAC Beachcroft
LegalNamed in an attributed testimonial from an Early Talent Assistant, covering training contracts, apprenticeships and internships.
View DAC Beachcroft guide →Cadent Gas
Energy & UtilitiesNamed in an attributed testimonial from a Talent Acquisition Manager describing bespoke screening built for its Future Talent 2025 intake.
View Cadent Gas guide →Kreston Reeves
Accountancy & Professional ServicesTestimonial naming critical thinking, verbal reasoning and a personality questionnaire, used to sift graduate and school leaver apprenticeship applicants.
View Kreston Reeves guide →Evelyn Partners
Wealth Management & AccountancyNamed in an attributed testimonial from an Early Careers Coordinator.
View Evelyn Partners guide →Vitality
Insurance & HealthNamed in an attributed testimonial from an Operational Support Lead.
View Vitality guide →Selco Builders Warehouse
Building Supplies & RetailTestimonial naming the Insights numerical, verbal and inductive tests by name. The clearest single statement of which products a customer runs.
STR Group
RecruitmentTestimonial naming the TPAQ-45 personality profile.
Pernod Ricard
Drinks & Consumer GoodsNamed in an attributed testimonial from an HR Business Partner, plus a homepage logo.
Monzo
Banking & FintechNamed in the "Trusted by Monzo, EDF, and more" strapline repeated across Test Partnership product pages. No detail and no case study, so treat as reported rather than confirmed.
View Monzo guide →EDF
Energy & UtilitiesNamed in the same product-page strapline as Monzo. Reported by the provider, with no test named.
GSR
Trading & Digital AssetsTestimonial from a Director, Head of Recruitment, naming the gamified MindmetriQ assessments specifically.
Every entry below comes from Test Partnership’s own site: an attributed testimonial, a homepage customer logo, or the strapline it repeats across its product pages. That is first-party evidence about its own customers, but it is undated and it is marketing, so confirm the current process for your role before preparing.
Practice by Test Type
Compare the same format across every provider that uses it, and see what each one is really measuring.
Numerical Reasoning Test Practice
Tables, charts, percentages and ratios against the clock. The most common screening test on graduate schemes and analyst roles.
Verbal Reasoning Test Practice
Short dense passages and true, false or cannot say answers. Judged on what the text supports, never on what you already know.
Logical Reasoning Test Practice
Shapes, sequences and flowcharts instead of words or numbers. Inductive, deductive, abstract, diagrammatic and spatial in one place.
Situational Judgement Test Practice
Realistic workplace scenarios scored against an employer competency framework. No maths, no time pressure, and a scoring key you can reason about.
Personality Test Practice for Job Applications
No right answers, which is exactly why they unsettle people. Learn the traits, the consistency checks and the formats before you sit one for real.
Ready to Start Practising?
Sign up for free to try 2 tests, or unlock all 9 Test Partnership simulations with unlimited practice and full feedback.