Average IQ by Country: The Rankings—and Why the Numbers Are Less Certain Than They Look

The United States appears at 101.04 in one current country ranking and 97.43 in a famous older dataset. Same country, very different number—the source matters as much as the decimal point.

There is no single authoritative worldwide census of average IQ by country. Current ranking pages mainly draw from two very different source families: large online-test datasets in which people across countries take the same test, and older compiled “national IQ” datasets assembled from many separate studies. The first improves test consistency but has self-selection and internet-access bias; the second tries to synthesize research samples but has much larger comparability and representativeness problems.

For context, an individual IQ test is typically scaled around a mean of 100 with a standard deviation of 15. See Cognitive Train’s guide to the average IQ score and our IQ score chart for how that scale works at the individual level.

Average IQ by Country 2026: Online Test Ranking

The most current large same-test ranking comes from the International IQ Test website. Its 2026 table is based on 1,212,714 people who took the same online IQ test during 2025. The test is scaled to a mean of 100 and a standard deviation of 15, and countries with fewer than 100 participants were excluded from the official ranking.

This solves one problem found in older national-IQ compilations: everyone took the same test. It creates another problem, though. These were people who chose to take an online IQ test, not citizens selected through a nationally representative probability sample. The source itself notes internet-access bias and large differences in participation between countries.

Territory note: the source’s official global ranking treats several territories under sovereign-country entries while also reporting their results separately. Hong Kong is listed separately at 107.73, Macao at 106.32, and Taiwan at 105.80. Some secondary ranking sites display those territory results as separate ranked entries, which is one reason their “highest IQ country” tables may look different.

Why You May See Different Country IQ Numbers Elsewhere

The older Lynn–Becker 2019 dataset is still widely repeated online. It was built from many separate cognitive-test studies and estimates, so it is a fundamentally different type of dataset from the 2025 same-test online sample.

The older dataset remains useful for understanding a large body of national-IQ research, but its decimal-level figures should not be mixed with the newer online ranking as though they came from one measurement system. The Lynn–Becker releases are available through the National IQ Dataset archive.

Why Researchers Disagree About Compiled National IQ Rankings

The strongest criticism is straightforward: many source samples were never designed to represent an entire country.

Rebecca Sear’s 2022 analysis of the 2019 dataset concluded that many inputs came from small or unrepresentative convenience samples, different cognitive tests, different age groups, and samples selected for special characteristics. She argued that the resulting national estimates are not sufficiently comparable for confident worldwide ranking.

Cross-country testing also raises questions about language, item design, and whether a test functions equivalently across cultures. Cognitive Train’s guide to cultural bias in IQ testing explains how researchers examine those problems.

A statement from the European Human Behaviour and Evolution Association similarly warned that national-IQ datasets fall short on data quality, validity, and transparency for reliable cross-cultural inference.

There is also a more moderate position. Russell Warne’s review of national mean IQ estimates argued against treating the entire idea as either perfectly valid or completely useless. His recommendation was to recognize the dataset’s shortcomings, improve measurement, and be much more careful about what conclusions the numbers can support.

Why the Rankings Can Still Correlate With School Achievement

Supporters of national-IQ estimates often point out that the scores correlate with international educational assessments. A study published in Intelligence found that national IQ estimates were strongly related to international measures of mathematics, science, and reading achievement.

That relationship is interesting, but it does not solve the sampling problem. Two country-level measures can correlate while one of them still contains substantial measurement error. School achievement itself also reflects education quality, family resources, health, language, test familiarity, and who remains enrolled in school.

A Better Cross-Country Comparison: PISA

If your goal is to compare cognitive-academic performance across countries using a standardized modern assessment, OECD PISA is much easier to interpret. PISA is not an IQ test; it tests 15-year-olds in mathematics, reading, and science using a coordinated international sampling framework.

In PISA 2022, Singapore scored highest in mathematics, reading, and science. Japan, Korea, Hong Kong, Macao, Chinese Taipei, Estonia, Canada, and several European systems also performed strongly, depending on the subject.

These results still do not tell you that one country is “more intelligent” than another. They tell you how sampled students performed on defined academic-cognitive tasks at a particular time.

Do Country Differences Mean Genetic Differences?

No country table can answer that question by itself. A group average is shaped by many influences that can differ across places and generations: schooling, nutrition, disease burden, socioeconomic conditions, language, test exposure, migration, and sampling all matter.

Heritability within a population also cannot be converted into a claim about why two national averages differ. Our guide to whether IQ is genetic or learned explains why those are separate questions.

Country Averages Are Not Permanent National Traits

Even when cognitive-test averages are measured carefully, they can move across generations. The Flynn effect is the best-known example. A large meta-analysis covering 271 samples and nearly four million participants found substantial worldwide gains on intelligence tests across much of the twentieth century, with different patterns across abilities and time periods.

That matters for country comparisons. A national average observed in one decade is not a permanent property of that population. Education, health, nutrition, family size, technology, test familiarity, and other social conditions can change. Migration also changes who is included in a national population.

There is another scale problem: people within the same country vary enormously. Two individuals from one country can differ far more from each other than two national averages differ. A country label therefore tells you very little about the ability of a particular person standing in front of you.

How Should You Read an “Average IQ by Country” Table?

Use it as a claim about a dataset, not as a precise fact about every person in a country. Ask which dataset produced the number, when the underlying tests were collected, whether the samples were nationally representative, whether the country was directly measured, and how large the uncertainty is.

The same caution applies to online testing. Cognitive Train’s guide to how accurate online IQ tests are explains why a clean-looking score can still depend heavily on test design and sampling.

See Where Your Own Score Falls

Country averages are poor tools for judging an individual. If you are curious about your own current reasoning profile, Cognitive Train’s IQ-Style Test provides an approximate estimate across several reasoning-related abilities.

For more targeted measures, browse Cognitive Train’s brain tests and cognitive assessments.

The Bottom Line

Average IQ by country rankings exist, but no single table gives an exact national measure of intelligence. The 2025 same-test online ranking offers excellent consistency of test format but a self-selected internet sample. The older Lynn–Becker figures draw from a broader research literature but face much larger problems of sample quality and cross-study comparability.

For a broader explanation of what IQ does and does not measure, read What Is IQ? or explore Cognitive Train’s Intelligence & IQ guides. You can also use our brain training and cognitive training tools to practice specific abilities directly.