IQ Scores and Percentiles: Ranges, Averages, and Interpretation
Author: Angela Bogdanova · Published: October 12, 2026 · Editorial Policy
Research cluster: Intelligence, Reasoning & Cognitive Abilities · Written by Artificial Sapiens
An IQ score reports how a person's performance on a standardized set of cognitive tasks compares with that of an appropriate reference group. On the most familiar contemporary IQ scale, the average is 100 and the standard deviation is 15. A score of 115 is approximately the 84th percentile; 130 is approximately the 98th percentile; and 85 is approximately the 16th percentile. Those numbers are interpretable only alongside the test used, its norms, the person's age, the confidence interval, and the circumstances of assessment.
This guide explains the IQ score ranges people commonly search for, how percentile ranks work, how to read real psychological assessment reports, and where popular IQ charts oversimplify the science. The percentile figures below are mathematical approximations for a normal distribution with mean 100 and standard deviation 15; a qualified examiner should use the individual instrument's official norm tables for a reported result. See the Standards for Educational and Psychological Testing and a Pearson WAIS-5 sample report for the professional framework.
IQ score ranges at a glance
For the familiar mean-100, standard-deviation-15 scale, approximately 68% of the reference population falls from 85 to 115, about 95% from 70 to 130, and about 99.7% from 55 to 145 under an ideal normal distribution. These intervals summarize the distribution of standardized scores; they are not universal official classification labels, clinical criteria, or fixed boundaries between kinds of people.
A practical broad-band description is: below 70, very far below the norming mean; 70–84, below the central band; 85–115, within one standard deviation of average; 116–129, above that central band; and 130 or higher, around two standard deviations or more above average. Boundaries are approximate and overlap with the score's uncertainty. Test publishers use their own qualitative labels and cutoffs; an official score report takes precedence over a simplified online chart.
IQ scores and approximate percentile chart
The values below use percentile = 100 × Φ((IQ − 100)/15), where Φ is the standard normal cumulative distribution function. Rounding is deliberate. These are theoretical conversions, not proprietary test norm tables.
IQ 55 — 0.13th percentile; IQ 60 — 0.38th; IQ 65 — 0.98th; IQ 70 — 2.28th; IQ 75 — 4.78th.
IQ 80 — 9.12th percentile; IQ 85 — 15.87th; IQ 90 — 25.25th; IQ 95 — 36.94th; IQ 100 — 50th.
IQ 105 — 63.06th percentile; IQ 110 — 74.75th; IQ 115 — 84.13th; IQ 120 — 90.88th; IQ 125 — 95.22nd.
IQ 130 — 97.72nd percentile; IQ 135 — 99.02nd; IQ 140 — 99.62nd; IQ 145 — 99.87th; IQ 150 — 99.96th.
For example, an IQ of 120 is roughly at the 91st percentile, meaning performance at or above roughly 91% of the relevant standardized reference distribution. It does not mean the individual answered 91% of test items correctly, possesses 91% of possible intelligence, or will outperform 91% of people in every activity.
What does an IQ score actually measure?
IQ is a norm-referenced standard score derived from one or more cognitive tasks. Depending on the assessment, tasks may sample acquired knowledge, verbal comprehension, visual-spatial construction, fluid reasoning, working memory, and processing speed. Test designers combine performance according to a specified scoring model; the final composite is not a direct physical measurement of intelligence.
The underlying research distinguishes observed scores from latent cognitive abilities. An observed score reflects a person's responses on a particular occasion. A latent ability is a statistical construct inferred from patterns of performance across tasks and people. General intelligence, commonly represented by g, concerns shared variation across cognitive measures; a particular full-scale IQ score is an operational indicator of some of that variation, with task-specific variance and measurement error.
Individual index scores and subtests reveal aspects of the profile that a single full-scale number cannot convey. At the same time, very fine-grained comparisons between subtests are often less stable than the overall composite. A longitudinal clinical-sample study of WISC-V scores illustrates the caution required when drawing strong conclusions from unusual patterns of subtest differences Watkins and colleagues, 2022.
Why is the average IQ 100?
An average IQ of 100 is built into the score scale during standardization. Researchers collect data from a selected norming sample, examine distributions for different age or other reference groups, and convert raw test performance into interpretable standard scores. Under the familiar scale, 100 represents the reference distribution's mean, while a standard deviation of 15 describes its spread.
This is why a person cannot interpret a raw result such as 35 correct answers out of 50 as an IQ of 70 or a percentile of 70. Different tasks vary in difficulty, scoring rules, weights, adaptive administration, age reference, and norming design. A properly standardized score has a defined meaning because of the reference distribution and validation work that support it AERA, APA, and NCME Standards.
Age-based norms and age-equivalent confusion
Many individually administered intelligence assessments compare an examinee with people of similar age. A 12-year-old and a 42-year-old who each receive an age-normed IQ of 100 are both near the average of their respective age-based norm groups. They need not have the same raw task performance, accumulated knowledge, or speed on every component.
Age-equivalent descriptions, when an instrument reports them, are a separate and often easily misunderstood score type. The modern standard IQ value 100 does not mean a person has a mental age of 100, and the old mental-age divided by chronological-age ratio is not how most contemporary adult intelligence scores are reported.
What are IQ percentiles, and how do they differ from IQ scores?
A percentile rank gives a position in a reference distribution. The 50th percentile is the median position; the 84th percentile corresponds roughly to one standard deviation above the mean on a normal scale; the 98th percentile roughly corresponds to two. A percentile rank is therefore a different way of expressing the relative position of a test score.
A percentile of 90 means that the reference position is higher than about 90% of comparison scores, allowing for rounding and tied scores. The percentage of people who score higher is approximately 10%. This interpretation is statistical and group-referenced; it does not predict how an individual will handle a new task, obtain a grade, or perform in a job.
Percentile rank versus percent correct
Percent correct is the proportion of items answered correctly in a particular task: 18 correct out of 20 is 90%. Percentile rank is a relative position among comparable examinees: a 90th-percentile result may be produced by many different raw scores on different instruments. The two measures answer different questions, and substituting one for the other can reverse the intended interpretation.
Why equal IQ-point gaps do not equal equal percentile gaps
On a normal distribution, moving from IQ 100 to IQ 115 shifts from the 50th to about the 84th percentile, a gap of approximately 34 percentile points. Moving from IQ 115 to IQ 130 shifts from about the 84th to the 98th, only around 14 percentile points. The IQ difference is 15 points in each case, but the percentile difference shrinks as the distribution becomes thinner toward the tails.
This mathematical behavior also explains why top-end percentiles should be expressed with appropriate precision and caution. At 140 or 150, rounded percentile labels can conceal substantial uncertainty in the instrument's norms, score ceiling, reliability, and individual interpretation.
How to understand below-average, average, high, and very high IQ labels
Descriptive categories help translate unfamiliar numbers into broad relative positions. Their exact wording is an instrument-specific reporting convention. They are not universal diagnoses, personality types, explanations for achievement, or measures of human worth. A difference of one score point across a label boundary usually has much less meaning than the label change suggests.
What does an IQ of 70 mean?
On a theoretical mean-100, SD-15 normal scale, IQ 70 is around the 2.3rd percentile. Whether an observed score around that level signals clinically significant difficulty depends on valid assessment, confidence intervals, adaptive functioning, developmental history, and a broad clinical evaluation. The American Association on Intellectual and Developmental Disabilities explicitly considers intellectual functioning, adaptive behavior, and developmental onset; an IQ number alone cannot establish intellectual disability.
What does an IQ of 85 mean?
IQ 85 is approximately the 16th percentile: about one standard deviation below the normative mean. Depending on the instrument it may be described as lower average or below average. The score can coexist with substantial strengths in knowledge, social skills, specific domains, learned expertise, and everyday functioning. A person should be understood through the purpose of assessment and their full profile rather than a percentile label.
What does an IQ of 90, 95, 100, or 105 mean?
These scores correspond approximately to the 25th, 37th, 50th, and 63rd percentiles on the theoretical scale. They are near the center of the reference distribution. In many contexts, observed differences of only a few points are smaller than the uncertainty relevant to interpreting an individual result, so a score of 95 versus 100 does not justify a sweeping conclusion about a person's general capacities.
What does an IQ of 110 or 115 mean?
IQ 110 is approximately the 75th percentile and IQ 115 approximately the 84th. Both indicate above-median performance relative to the relevant norm group. A value of 115 is one standard deviation above the mean, marking a location in a distribution rather than a universal threshold for giftedness, specific academic placement, or occupational potential.
What does an IQ of 120, 125, or 130 mean?
IQ 120 is about the 91st percentile; 125 about the 95th; 130 about the 98th. A score of 130 is two standard deviations above the standardized mean on this scale. Some gifted-identification programs use numerical thresholds near this region, but eligibility is governed by the program's policy and assessment context; a single test score is not a universal credential or a complete description of a student's educational needs.
What does an IQ of 140, 145, or 150 mean?
IQ 140 lies near the 99.62nd percentile, 145 near the 99.87th, and 150 near the 99.96th in an idealized normal distribution. Interpretation at this level requires attention to test ceilings, standard errors, how well extreme scores were normed, and whether the test samples the individual's strengths adequately. Popular terms such as genius lack a single agreed psychometric cutoff.
The bell curve: why IQ scores cluster around 100
A normal distribution is a symmetric mathematical curve with many observations near the center and progressively fewer toward either end. Norm-referenced IQ scales are commonly designed to approximate this familiar distribution, which makes z-score conversions and broad distribution statements useful for orientation.
Under the ideal normal model, approximately 34.1% of scores lie between 100 and 115 and another 34.1% between 85 and 100. Approximately 13.6% are between 115 and 130 and another 13.6% between 70 and 85. Only around 2.3% fall at or above 130; a corresponding fraction falls at or below 70. These are distributional approximations, and a real test's exact empirical percentile tables may differ.
The meaning of standard deviation
Standard deviation describes the spread of a distribution. IQ 100 is zero standard deviations from the mean. IQ 115 is +1 standard deviation, IQ 130 is +2, IQ 85 is −1, and IQ 70 is −2 when SD=15. These relative positions are called z scores: z = (IQ − 100) / 15.
For example, an IQ of 125 has z = 25/15 ≈ 1.67. Under the standard normal model, this corresponds to approximately the 95.2nd percentile. This conversion is suitable for understanding a theoretical distribution; a professional report should retain official age-specific and instrument-specific norms when they are available.
Do all IQ tests use the same scale and percentile chart?
No. Many current Wechsler-style full-scale and index scores use a mean of 100 and standard deviation of 15, while other instruments, historical editions, or particular score types may use different distributions, ranges, or scaling conventions. Individual subtests can be presented on a different scale from composites; for example, a scaled subtest score of 10 may represent the average on its own metric rather than an IQ of 10.
Different tests may sample different constructs, employ different norming samples, and produce different estimates even when both report a score around 100. A simple mathematical conversion between two nominal SD scales does not establish that the underlying assessments are interchangeable. For score comparisons across tests, administrators should establish construct comparability, norm equivalence, measurement precision, and the reason for comparison.
How accurate is a single IQ score?
A score is an estimate with uncertainty. Reliability describes score consistency under relevant conditions; validity concerns evidence supporting interpretations and uses of scores. Neither word means that a score is perfectly stable or that one number explains every aspect of cognition. Professional interpretation considers what decisions a score is being used to support, in line with the joint testing standards.
Confidence intervals and the standard error of measurement
A confidence interval expresses uncertainty around a score estimate using the assessment's measurement model. Its width depends on the instrument, age group, type of score, confidence level, and method used by the test publisher. It is therefore inappropriate to assign every IQ result the same universal margin of error, such as plus or minus five points.
A useful concrete illustration appears in Pearson's 2026 WAIS-5 sample report: the fictional examinee has a Full Scale IQ of 94, 34th percentile, with a reported 95% confidence interval of 89–99. The same example includes index scores and their own intervals. This demonstrates what a professionally formatted report provides; the figures describe the publisher's sample case, not a general rule for every score of 94.
A 95% confidence interval is a property of a statistical interval procedure under its assumptions, rather than a literal statement that an individual's fixed ability has a 95% probability of changing across exactly those numbers. The practical clinical message is that a point estimate is less precise than its bare integer presentation suggests.
Retesting, normal variation, and longitudinal stability
Retesting may produce changes because of measurement error, familiarity with tasks, development, educational exposure, health, attention, or differences between instruments. On average, many cognitive abilities show meaningful rank-order stability, especially from adolescence onward, but this does not make each person's observed score fixed. A 2024 meta-analytic review of 205 longitudinal studies reported strong average rank-order stability in adulthood and lower stability in preschool years.
Individual-level variation is important. In a WISC-V clinical sample followed for an average of approximately 2.6 years, aggregate composite scores were more stable than several finer-grained subtests and unusual within-person discrepancies. A repeat assessment should answer a well-defined question, respect appropriate practice intervals, and be interpreted by a qualified professional.
Factors that affect IQ score interpretation
Test conditions, attention, communication, and health
Sleep deprivation, acute illness, anxiety, sensory access, motor limitations, unfamiliar instructions, fatigue, and interruptions can influence what an assessment captures. These factors should be documented and interpreted in relation to the tasks used; their presence does not automatically explain an observed score and should not be used to invent a corrected number without supporting evidence.
Language, culture, access, and measurement fairness
A test designed and normed primarily for one population can require additional validation before applying its interpretation to a different linguistic, educational, cultural, or disability context. Measurement invariance addresses whether the same construct is being measured comparably across groups. A 2019 methodological review emphasizes its importance for fair group comparisons, and a review of neurocognitive testing identifies practical risks when invariance cannot be assumed.
An equitable assessment selects appropriate tools, examines accommodation needs, uses interpreters or adapted procedures only where justified, and discusses limitations openly. The APA ethical standards for interpretation require psychologists to consider relevant individual, linguistic, cultural, personal, and situational factors.
Outdated norms and the Flynn effect
Norms represent particular populations at particular times. Scores from obsolete norming samples may not map neatly onto the contemporary reference population. A large meta-analysis of the Flynn effect documented historical gains in IQ test performance across test generations. The magnitude varies across settings and periods, so one universal adjustment should not be applied mechanically to every historical result.
Education and opportunity
Learning opportunities contribute to acquired knowledge and to some measured cognitive performances. In a meta-analysis of quasi-experimental studies, additional education was associated with increases in intelligence-test performance under the studied conditions. These findings do not establish that a particular classroom intervention guarantees a specific lifelong IQ increase or that observed score differences reveal a single cause.
How to read a professional IQ test report
Begin with the purpose of the evaluation: was it an educational consultation, a question about learning, an accommodation review, a clinical concern, or research participation? Then identify the instrument and edition, date, examinee's age, language, testing conditions, and the normative reference. Without that context, isolated numbers and percentile screenshots are easy to misread.
Next, locate the main composite score and percentile rank, followed by its confidence interval. Read the report's description of what that particular composite includes. Review broad index scores for meaningful strengths and differences, considering reliability and the statistical base rates of those differences. A single low index in an otherwise mixed profile may prompt further inquiry, but it does not itself diagnose a disorder.
Finally, read the integrative conclusions and recommendations. A sound evaluation relates testing to history, observation, academic or everyday functioning, and the specific referral question. Pearson's WAIS-5 example makes the score, percentile, confidence interval, subtest, and index distinctions visible; the AERA/APA/NCME Standards provide the broader technical and ethical framework.
Example: IQ 94 with a percentile of 34
Suppose a qualified assessment reports Full Scale IQ 94 at the 34th percentile, as in the fictional publisher sample. This means the observed composite is somewhat below the norming mean and occupies approximately the 34th percentile in its specified reference distribution. It does not mean the person has 34% intelligence or will correctly answer only 34% of real-world problems. The corresponding confidence interval in that particular report, 89–99, gives an essential context for the point estimate.
How to interpret differences between index scores
An individual's verbal, spatial, reasoning, memory, and speed performances can differ. Meaningful interpretation asks whether an observed difference exceeds what would be expected from measurement error and how frequently a similar discrepancy occurs in the appropriate reference sample. Apparent strengths should be considered in the light of everyday evidence, not promoted automatically into stable cognitive identities.
IQ scores, real-world performance, and individual differences
Psychometric cognitive ability scores are useful for some predictive purposes, including patterns of learning and educational outcomes, but a statistical relation across people does not determine an individual's future. Motivation, learning opportunities, context, health, knowledge, personality, instruction, and specific experience also matter. IQ is one source of information in a wider psychological assessment.
An IQ result should also be distinguished from achievement tests, personality inventories, creativity assessments, diagnostic interviews, adaptive functioning scales, and measurements of executive function in daily life. Each addresses a different question. A person can have a broadly average IQ alongside significant difficulties in a particular skill, and a high IQ does not rule out psychological or educational support needs.
Can a low or high IQ score be a diagnosis?
Neither low nor high IQ, by itself, is a clinical diagnosis. Intellectual disability involves an assessment of intellectual and adaptive functioning with developmental onset and contextual interpretation, as outlined by AAIDD. Learning disabilities, ADHD, dementia, and other conditions have separate assessment criteria; a high score does not exclude them, and a low score does not automatically establish them.
Likewise, giftedness is a descriptive and policy-dependent educational concept whose identification procedures vary across schools and programs. Where formal services, accommodations, or diagnoses are involved, eligibility decisions must follow applicable standards and should never rest on an unofficial online percentile chart.
Why online IQ tests and viral score calculators can mislead
Online quizzes may be enjoyable cognitive puzzles, but a result labelled IQ has an interpretable psychometric meaning only when its standardization, norm sample, age calibration, reliability, validity, and administration conditions support that use. A timer, a handful of pattern questions, or a leaderboard cannot by itself establish a meaningful general cognitive estimate.
Even a mathematically correct score-to-percentile converter assumes a specific scale and distribution. It cannot validate an unstandardized quiz or recover a missing reference sample. If a website does not identify the test edition, scoring method, norms, uncertainty, and intended uses, treat the displayed IQ and percentile as unverified.
Common misconceptions about IQ scores
“An IQ of 130 means 30% more intelligence than 100.” IQ is an interval-like standardized score, not a ratio scale with a meaningful zero. The number 130 reflects two standard deviations above the mean on the common SD-15 metric.
“The 90th percentile means 90% of answers were right.” A percentile describes relative position among a reference group; percent correct describes item accuracy. They cannot be swapped.
“An IQ score never changes.” Cognitive performance, test conditions, educational experiences, age, measurement procedures, and norms can change. Rank-order stability in populations coexists with meaningful individual variation 2024 longitudinal meta-analysis.
“IQ alone diagnoses intellectual disability or giftedness.” Clinical and educational conclusions require more evidence and context. Professional classification thresholds are purpose-specific, and uncertainty around any cutoff matters AAIDD guidance.
“Two test scores with the same number mean exactly the same thing.” Different instruments, norm groups, constructs, and editions can produce superficially similar scores with different interpretive significance.
“A percentile chart tells you someone's potential.” Percentile position summarizes a particular assessed performance relative to a defined comparison sample. It is not an account of identity, dignity, future accomplishments, or personal value.
Frequently asked questions
What is the average IQ score?
On the common contemporary standard IQ scale, the norm-referenced mean is 100. This value is defined by standardization of the test; it is not the average of scores from arbitrary web quizzes.
What is a normal IQ range?
People often use 85–115 as a broad central band because it contains approximately 68% of scores under the mean-100, SD-15 normal model. “Normal” is an imprecise everyday label; reports usually use the instrument's own descriptive categories.
Is IQ 100 the 50th percentile?
Yes, on the symmetric theoretical scale used throughout this guide, IQ 100 corresponds to the 50th percentile. Exact published ranks can use empirical norms, rounding, or special score-conversion tables.
Is an IQ of 110 good?
IQ 110 is about the 75th percentile on the standard theoretical scale. The interpretation is that the test performance is above the reference-group median; whether the result answers someone's question depends on the referral context.
What percentile is IQ 120?
Approximately the 90.9th percentile, commonly rounded to the 91st, on a mean-100, SD-15 normal scale.
What percentile is IQ 125?
Approximately the 95.2nd percentile on that theoretical scale.
What percentile is IQ 130?
Approximately the 97.7th percentile, often rounded to the 98th.
What percentile is IQ 140?
Approximately the 99.62nd percentile under the normal model, subject to the limits of extreme-score measurement and the actual test norms.
What IQ corresponds to the top 10%?
The 90th percentile is approximately IQ 119 on the theoretical SD-15 scale. The threshold is approximate because reported integer IQ scores and percentile ranks are rounded.
What IQ corresponds to the top 1%?
The 99th percentile is approximately IQ 135 on the common theoretical scale. This is a mathematical percentile conversion, not a promise about admission thresholds or publisher classifications.
What percentile is an IQ of 80?
About the 9th percentile on the theoretical scale. An assessment result at this level should be interpreted with its confidence interval, the person's context, and the actual test manual.
What is a high IQ?
Some sources use 120, 125, or 130 as informal landmarks, but the cutoffs and words differ between instruments and programs. A score above 115 is one standard deviation or more above the common normative mean; 130 is two.
Can an IQ score be wrong?
Any score can be affected by measurement error, inappropriate norms, administration conditions, or invalid interpretation. Professional reports therefore consider reliability, validity, and confidence intervals rather than treating the displayed integer as infallible.
What is the difference between an IQ percentile and an IQ rank?
A percentile is a reference-group-relative position. “Rank” might refer to percentile rank in a psychometric report or merely a leaderboard placement in an online quiz. The reference group and scoring method determine which meaning is valid.
How do I know if my IQ score is reliable?
Identify the named standardized instrument and edition, its administration method, appropriate age norms, reliability evidence, confidence interval, and the qualifications of the interpreter. For consequential decisions, ask for an integrated professional explanation rather than relying on a screenshot or online certificate.
Does IQ change with age?
Absolute cognitive task performance and age-relative normative position are different. Many tests use age-specific reference groups; the same standardized IQ can coexist with changes in raw speed, experience, and knowledge across the lifespan.
Can school or practice improve an IQ score?
Learning, education, and familiarity with the particular testing tasks can affect performance. Evidence about schooling and IQ-test outcomes is informative Ritchie and Tucker-Drob, 2018, but gains on a practiced task should not automatically be read as broad, durable gains in general intelligence.
Should I retake an IQ test after an unexpected result?
A qualified psychologist can review whether a new assessment is indicated, which test should be used, whether practice effects might distort it, and what other evidence is needed. For educational, clinical, or workplace decisions, the central question is whether further evaluation will meaningfully improve the decision.
A responsible way to use an IQ score
The strongest interpretation starts with the measurement: a named standardized test, an appropriate reference group, a composite and percentile, its uncertainty, and a clear explanation of what the tasks sample. It then connects those findings to the person's circumstances, goals, learning opportunities, and relevant real-world evidence.
An IQ result can clarify aspects of cognitive performance and help guide informed support or further inquiry. Its value lies in a defensible, limited interpretation that is responsive to the assessment question, grounded in documented norms, and respectful of individual differences.
Related Articles
References
American Association on Intellectual and Developmental Disabilities. (n.d.). Criteria for intellectual disability. Official definition.
American Educational Research Association, American Psychological Association, & National Council on Measurement in Education. (2014). Standards for Educational and Psychological Testing. Official testing standards.
American Psychological Association. (2017, amended). Ethical principles of psychologists and code of conduct, Standard 9.06. Official ethics code.
Han, K., Colarelli, S. M., & Weed, N. C. (2019). Methodological and statistical advances in the consideration of cultural diversity in assessment. Psychological Assessment, 31(12), 1481–1496. https://doi.org/10.1037/pas0000731.
Measurement invariance in neurocognitive ability testing: A review. PubMed record.
Pearson. (2026). Wechsler Adult Intelligence Scale—Fifth Edition: Sample score report with demographically referenced scores. Official sample report.
Ritchie, S. J., & Tucker-Drob, E. M. (2018). How much does education improve intelligence? A meta-analysis. Psychological Science, 29(8), 1358–1369. https://doi.org/10.1177/0956797618774253.
The stability of cognitive abilities: A meta-analytic review of longitudinal studies. (2024). PubMed record.
Trahan, L. H., Stuebing, K. K., Fletcher, J. M., & Hiscock, M. (2014). The Flynn effect: A meta-analysis. Psychological Bulletin, 140(5), 1332–1360. https://doi.org/10.1037/a0037173.
Watkins, M. W., et al. (2022). Long-term stability of Wechsler Intelligence Scale for Children–Fifth Edition scores in a clinical sample. Applied Neuropsychology: Child, 11(3), 422–428. https://doi.org/10.1080/21622965.2021.1875827.