What Is a Good IQ Score? IQ Scale and Percentiles Explained

What Is a Good IQ Score? Start with What the Number Means

An IQ score can feel like a single number that explains everything about intelligence. In reality, its meaning depends on which assessment was used, who it was standardized on, how it was administered and what the score is being compared against. On many modern, professionally standardized tests, the average is set to 100 and the standard deviation is 15. That is the starting point for interpreting 85, 100, 115, 130 or 145.

The short answer: 100 is average on that scale; a higher or lower number describes performance relative to the test's norm group. There is no universal cutoff for being a “good” or valuable person. Nor does a brief online reasoning quiz automatically generate an official IQ or a diagnosis.

IQ Score Chart: Mean, Standard Deviation and Percentiles

Modern deviation IQ scores compare a person's performance with a specified normative population. The mean is typically 100; an SD of 15 means a score of 115 is one standard deviation above that mean. It does not mean someone answered 115 questions correctly or is “15% smarter.” The APA definition of deviation IQ explains this difference.

The following figures are approximate mathematical percentiles for a normal distribution with mean 100 and SD 15. They are not measured ReflexBenchmark visitor statistics, and test publishers may use different scoring conventions or norm tables.

Percentiles rounded to whole numbers except at 145. These follow the ideal normal-distribution model; official reports should use the specific test's norms, score intervals and administration rules.

Is 100, 115 or 130 a “Good” IQ?

100 is the chosen midpoint on many standardized scales and therefore is not a failure. 115 is above the test's norm-group average. 130 is about the 98th percentile on the familiar SD-15 model; some educational programs use cutoffs around this range, but eligibility rules can differ. A person's strengths can be uneven across verbal understanding, pattern reasoning, memory and speed, so a single number can conceal important distinctions.

Membership in Mensa is based on the upper 2% on an approved, properly administered test, not merely entering “130” into a generic online quiz. Mensa notes that different test instruments can have different score scales and that an online test does not qualify for admission.

What Do Standardized IQ Tests Actually Measure?

A professionally administered cognitive battery usually samples several kinds of tasks. Depending on the instrument and version, these can include verbal comprehension, visual or fluid reasoning, working memory and processing speed. It is inaccurate to say every modern test uses the same four subtests: the exact index structure depends on the edition and the publisher.

• Verbal comprehension: understanding words, concepts and learned language.
• Fluid and visual reasoning: identifying unfamiliar relationships or patterns, often with abstract designs.
• Working memory: holding and manipulating small amounts of information while completing a task.
• Processing speed: performing relatively simple visual or symbolic tasks efficiently under time limits.

These abilities overlap but are not identical. A person who completes a fast visual-search task may not get the same standing on a vocabulary or abstract-reasoning task. Likewise, completing five good reaction-time trials does not produce a reliable full-scale IQ score.

For more detail on the distinction between novel reasoning and learned knowledge, see our Fluid vs Crystallized Intelligence guide. Our What Is Working Memory? article explains why short-term memory measures are related to, but not synonymous with, general reasoning.

Why an IQ Percentile Is Not a Percentage Correct

A percentile is a relative rank within a norm group. Someone at the 84th percentile performed at or above roughly 84% of that group on the relevant normed metric. It is not an 84% quiz accuracy score. Two individuals answering the same number of puzzles correctly may even receive different standardized scores if norms vary by age group, edition or administration procedure.

This distinction also explains why popular “Top 1%” graphics are risky when no valid population comparison has been collected. A website may know your raw puzzle score but not have a representative sample or enough comparable attempts to estimate a true percentile. For ReflexBenchmark, on-site recorded-attempt rankings, where offered, are rankings of site participants—not standardized general-population IQ percentiles.

Why Low or High IQ Scores Need Context

An IQ score does not capture all of a person's strengths or weaknesses. Tests have measurement uncertainty, and factors such as language, disability accommodations, education, fatigue and cultural familiarity may affect how results should be interpreted. A professional report can consider confidence intervals, different index scores and relevant personal history rather than treating a single number as exact.

Importantly, an IQ below approximately 70 or 75 is not by itself a diagnosis of intellectual disability. The American Psychiatric Association explains that assessment involves both intellectual functioning and adaptive functioning, including conceptual, social and practical everyday skills. Clinical conclusions belong to qualified practitioners, not an online score chart.

At the higher end, labels like “gifted” can be defined differently by schools and organizations. A standardized high score may indicate strong performance on certain tested abilities, but it does not guarantee success, wisdom or creativity in every real-world situation. The appropriate interpretation depends on the question the assessment was intended to answer.

Does IQ Predict Academic or Work Success?

Research often finds a statistical relationship between cognitive measures and outcomes such as some educational or workplace performance indicators. However, an association is not destiny. Different jobs, measures and samples can yield different results. One contemporary 2024 meta-analysis covering 153 samples found a corrected mean general cognitive ability–job performance relationship of approximately 0.22 in 21st-century data—lower than some older, widely repeated estimates.

The finding does not mean IQ “explains 22% of work success.” A correlation and a percentage of explained variance are different. It also says little about a specific person's motivation, training opportunities, interpersonal skills, circumstances or professional experience. That's why the original article's claims that a particular score is “typical of PhDs” or “typical of college graduates” have been removed: no defensible norm source was provided for those labels.

Can IQ Scores Change? What the Flynn Effect Shows

Standardized scores may vary across repeated testing and across the lifespan. Practice with a specific format, changes in health, development, test conditions and statistical measurement error can all play roles. This does not mean one short session can reliably improve general intelligence; improving a familiar puzzle score is not automatically the same thing.

The Flynn effect refers to historical shifts, often upward, in intelligence-test performance across generations. A 2015 meta-analysis by Pietschnig and Voracek synthesized 271 independent samples from multiple countries and reported that changes differed by cognitive domain, country and historical period. Another meta-analysis of normative differences reported an overall estimate near 2.3 standard-score points per decade across its selected studies. Neither supports claiming a constant gain for every country or every generation.

Because population performance and test content can change, publishers periodically revise norms and editions. A test result should therefore be interpreted against the population and edition it actually used, rather than with a generic internet table assembled from unrelated tests.

Can a Free Online Quiz Tell You Your IQ?

Online puzzles can be engaging ways to explore reasoning styles, but the phrase “IQ test” alone does not establish standardized scoring. A valid clinical estimate requires appropriate norms, psychometric reliability, validity evidence, administration conditions and interpretation. A small website quiz usually cannot reproduce all of those features.

On ReflexBenchmark, our Mini IQ-style Reasoning Quiz contains 20 practice questions and gives a raw score out of 20. It does not calculate a diagnostic IQ range. The Pattern Recognition Test and Digit Span are also informal browser exercises, not clinical substitutes. For more context before using any online test, read Free IQ Test Online: What to Expect.

How to Interpret Your Own IQ Report Responsibly

1. Identify the test and edition. A score means little without knowing how it was generated.
2. Confirm the comparison group. Check whether the norm is appropriate for the person's age, language and other relevant conditions.
3. Read the interval and subtests. Index scores and uncertainty can add detail that a single overall number misses.
4. Avoid self-diagnosis. If the assessment affects school support, clinical care or accommodations, seek qualified interpretation.
5. Remember the bigger picture. Intellectual tests are one source of information, not a verdict on a person's worth or future.

Sources and Further Reading

1. American Psychological Association. IQ and Deviation IQ, APA Dictionary of Psychology.
2. Mensa International. Getting Your IQ Tested: FAQs. Approved-test percentile admission requirements and differing test scales.
3. American Psychiatric Association. Intellectual Disability. The role of adaptive functioning alongside cognitive testing.
4. Sackett, P. R. et al. (2024). A Contemporary Look at the Relationship Between General Cognitive Ability and Job Performance. Journal of Applied Psychology.
5. Pietschnig, J., & Voracek, M. (2015). One Century of Global IQ Gains: A Formal Meta-Analysis of the Flynn Effect. Perspectives on Psychological Science.
6. Trahan, L. H. et al. (2014). The Flynn Effect: A Meta-analysis. Psychological Bulletin.