Find out where your reasoning lands on the scale. This short test looks at four different areas of thinking, and it takes about four minutes to finish.
Thousands of people have taken this one. Most of them are surprised by the result, and usually not in the direction they expected.
One thing to be clear about before you start: this is a quick estimate for fun and practice, not a clinical assessment. A proper IQ evaluation is run by a psychologist and takes an hour or more. What this test gives you is a decent read on how your reasoning is working today.
Answer the eight questions below carefully. There is no timer and no penalty for a wrong answer, so read each one twice.
What each question was actually testing
Every question above targets a specific piece of reasoning. Here is what each one was looking for, and why it belongs in a test like this.
Question 1: sorting things into groups
Picking the item that does not belong is one of the oldest tasks in cognitive testing, and one of the most basic. To answer it you have to hold five items in mind, find the rule that four of them share, and check the fifth against it. Apple, banana, grape and strawberry are fruit. Carrot is a vegetable.
You use this constantly without noticing. Sorting laundry, ranking your inbox, choosing the right word for a sentence. All of it runs on the same machinery. At a higher level, the same skill is what lets people build classification systems, spot the outlier in a spreadsheet and notice that one number in a report does not fit the others.
Question 2: analogies and the shape of a relationship
A book has an author. A painting has a painter. The trap here is the museum, because paintings and museums do go together, just not in the same way. Analogical reasoning is about the relationship, not the association.
This is the skill people lean on when they learn something new by comparing it to something familiar. Explaining electricity as water in a pipe works because the relationship transfers. Analogy questions show up on almost every standardized test for the same reason: they are hard to answer by memory and easy to answer by thinking.
Question 3: keeping a chain of facts in order
Mary, John, Peter, Anna. Four names, three comparisons, and no picture to lean on. This is transitive reasoning, and what it really measures is working memory. You have to build the order in your head and keep it stable while the next sentence arrives.
People who get this wrong usually get it wrong at the last step, because they stopped tracking one of the names. That is the interesting part: the logic is trivial, the load is not. It is the same load you carry when someone gives you directions with four turns in them.
Question 4: syllogisms with nonsense words
Bloops, Razzies and Lazzies are made up on purpose. Strip out the meaning and the only thing left is the form of the argument. If every A is a B, and every B is a C, then every A is a C. That is a valid syllogism and the answer is true.
Nonsense words are used deliberately in logic testing because real words bring opinions with them. People asked to judge an argument about politics or money will accept a broken argument that agrees with them. Replace the nouns with Bloops and that bias has nothing to grab onto.
Question 5: number sequences
2, 6, 18, 54. Each number is three times the one before it, so the next one is 162. The most common wrong answer is 108, which comes from doubling the last number instead of finding the rule that generated all of them.
Sequence questions test inductive reasoning: you see examples and work backwards to the rule. That is the same thing a scientist does with data and a trader does with a price chart. The useful habit here is to look at what happens between the numbers rather than at the numbers themselves.
Question 6: completing a visual matrix
The grid shifts one position on every row, so each shape moves left and wraps around to the end. Following that, the missing cell is a triangle. This is the format used in Raven progressive matrices, which is probably the most respected nonverbal intelligence test in existence.
Matrix questions are valued because they need almost no language and no schooling. Someone who never learned to read can solve them. That makes them one of the fairest ways to compare reasoning across very different backgrounds, which is why they turn up in research far more often than word puzzles do.
Question 7: the clock at 3:15
Almost everyone answers zero, because at 3:15 both hands look like they are on the 3. They are not. The minute hand is exactly on the 3, but the hour hand has already crept a quarter of the way toward the 4. A full hour covers 30 degrees, so a quarter of an hour moves the hour hand 7.5 degrees.
This question is in the test because it punishes the fast answer. Every part of it is easy arithmetic. What it asks for is the willingness to check a picture you think you already understand, which is a genuinely rare habit.
Question 8: the 17 sheep
All but 9 died. So 9 survived. The number 17 is there purely to tempt you into subtracting, and most people who answer 8 have done the arithmetic perfectly on a question that never asked for arithmetic.
Psychologists call this a cognitive reflection problem. It measures whether you can catch your own first instinct before you act on it. That skill matters more outside a test than inside one: it is the difference between reading a contract and skimming it, or checking a price and assuming it.
Intelligence is bigger than one number
We live in a world that likes to reduce intelligence to a single score. The score is real and it does measure something. It just measures a narrow slice of what people actually do with their minds.
Some people solve equations easily. Others read a room in seconds, or write music that makes strangers cry, or find a way through a problem that everybody else called impossible. Those are all forms of intelligence and none of them show up in eight multiple choice questions.
Howard Gardner argued for at least eight kinds: logical, linguistic, musical, spatial, bodily, interpersonal, intrapersonal and naturalist. You can disagree with his framework and still take the point. A low score on a reasoning test tells you about reasoning on that test. It does not tell you what you are capable of.
How to actually get better at this
The brain-training industry has oversold itself for years, and the research is blunt about it: getting better at a puzzle mostly makes you better at that puzzle. That said, a few habits do carry over, and none of them cost money.
- Read things that are slightly too hard for you, and finish them.
- Learn a game with real rules. Chess, bridge and go all force you to think several steps out.
- Argue properly with people who disagree with you, and let them finish their sentences.
- Sleep. Reasoning is the first thing to break down when you are short on it.
- Slow down on the questions that look obvious. That single habit is worth points on any test.
The last one is worth repeating because it is the cheapest. On this test, questions 7 and 8 both punish speed and nothing else. People who reread them get both right. People who trust the first answer that feels right get both wrong, and lose 12 points doing it.
Where questions like these come from
The first practical intelligence test was built by Alfred Binet in Paris in 1905, and he built it for a narrow reason: the French government wanted to find children who needed extra help at school. Binet was clear that his scale measured current performance, not fixed potential, and he worried in print about people treating the number as a permanent label.
The formula that gave us the term intelligence quotient came later, from William Stern, and the modern scales dropped it decades ago. Today a score is worked out by comparing you with everyone else who took the same test, which is why 100 always sits in the middle no matter how the questions change.
The question types on this page are the ones that survived a century of that work: analogies, sequences, matrices and short logic problems. They stuck around because they keep predicting performance on other reasoning tasks, which is a low bar in some ways and a hard one to clear in others.
Questions people ask about this test
Why is the maximum score 48 and not 8?
Because the questions are not worth the same. A fully correct answer is worth 6 points, and a few of the wrong options are worth 1 point because they show you understood the question and slipped on one step. Answering everything correctly gives you 48.
Why does the result show a range instead of a single number?
Eight questions cannot pin down a precise figure, and any site that gives you one is making it up. A range is the honest answer at this length. Real assessments use dozens of items per area for exactly this reason.
Is a score of 100 good or bad?
It is average, and average is where most people are by design. The scale is built so that 100 sits at the middle of the population. About two thirds of everyone falls between 85 and 115.
Does a bad night of sleep really change the result?
Yes, and more than people expect. Working memory is the first thing to suffer when you are tired, and questions 3 and 5 lean on it heavily. If you took the test half asleep, the score is about your night, not about you.
Can I share my result?
Send the page to whoever you want. Nothing is stored, so there is no link with your score baked into it. The other person takes the same eight questions from scratch.
Before you take the score too seriously
This test is entertainment and mental exercise. It does not replace a psychological assessment. Standardized instruments such as the Wechsler Adult Intelligence Scale and the Raven progressive matrices are given by qualified professionals under controlled conditions, and they produce results that are far more precise than anything eight questions on a phone can offer.
Whatever you scored, remember that intelligence is not fixed. The brain keeps changing throughout life. Reading, learning new things, staying curious and taking on problems that are a bit beyond you are the things that keep a mind sharp, and they work at every age.
If the result genuinely made you curious about your own cognitive profile, a neuropsychologist can run a proper evaluation and give you a real picture rather than an estimate. For everyone else: take it again in a few months, and see whether reading the questions twice changes anything.
