— Guide
The major IQ tests, compared
"IQ test" is a category, not a specific instrument. Different tests measure different cognitive subsets, score on different scales, and serve different purposes. Here are the major ones you'll encounter — clinical, supervised, and modern online.
Wechsler scales (WAIS-IV, WISC-V)
Format: Supervised, individually administered by a licensed psychologist. 60–90 minutes. Mean 100, SD 15.
What it measures: Verbal comprehension, perceptual reasoning, working memory, processing speed. Returns a Full Scale IQ plus four index scores.
Best for: Clinical assessment, educational diagnostics, formal cognitive evaluation. The default gold-standard test in most clinical settings.
Cost: $500–$2,000 administered privately.
Stanford-Binet (SB5)
Format: Supervised, individually administered. 45–75 minutes. Mean 100, SD 16.
What it measures: Fluid reasoning, knowledge, quantitative reasoning, visual-spatial processing, working memory. Strong upper ceiling, good for measuring high IQ.
Best for: Gifted assessment in children, high-IQ evaluation in adults. Mensa accepts SB5 scores at 132+.
Cost: $500–$1,500 administered privately.
Raven's Progressive Matrices
Format: Pure pattern-matching, no verbal content. Can be administered in groups, supervised. 40–60 minutes. Originally raw-score, modern norms convert to standard IQ.
What it measures: Fluid intelligence and abstract reasoning only — no crystallised knowledge component.
Best for: Cross-cultural assessment (no language needed), pure-reasoning measurement, research. Comes in Standard (SPM) and Advanced (APM) versions.
Mensa accepts: 98th percentile equivalents.
Cattell Culture Fair III
Format: Supervised group test. ~30 minutes. Mean 100, SD 24 (note the wider SD).
What it measures: Fluid intelligence, designed to minimise cultural and language confounds.
Best for: Cross-cultural settings, the historical Mensa admission test in the UK. The SD=24 means a "Cattell IQ" of 148 corresponds to a Wechsler 130.
Modern online IRT tests (including Core Brain)
Format: Unsupervised, web-based. 15–30 minutes. Mean 100, SD 15 (Wechsler convention).
What they measure: Fluid reasoning across multiple axes — pattern, spatial, logic, verbal, numeric, matrix. Uses item response theory (IRT) to estimate latent ability θ from your specific response pattern.
Best for: Calibrated self-assessment, getting a realistic estimate before paying for a supervised test, exploring cognitive strengths and weaknesses across axes.
Not for: Mensa admission, clinical diagnosis, employment decisions, academic placement.
The clickbait online quiz (avoid)
Format: Unsupervised, web-based. 5–10 minutes. No published methodology, no calibration disclosure.
What it measures: Almost nothing useful. These quizzes are designed to hand out inflated, flattering scores to maximise sharing.
Best for: Entertainment, not assessment. The output is closer to a horoscope than a measurement.
How to spot a real test from a fake one
- Disclosed methodology. A real test tells you it uses IRT or a normed reference distribution. Fake tests don't mention any of this.
- Confidence intervals. Any genuine IQ estimate comes with measurement error. If a test gives you a single number with no uncertainty band, the number isn't a real estimate.
- Difficulty progression. Real tests adapt difficulty — easier items first, harder items as you progress, or stratified by axis. Fake tests pile in the same easy items for everyone.
- No celebrity comparisons. "You scored higher than Einstein!" is a fake-test signature. Einstein never took a modern IQ test and his quoted scores are retrospective fiction.
Take a calibrated test
Core Brain runs a 3PL Bayesian IRT estimator with items calibrated to the Wechsler distribution. Twenty minutes, six cognitive axes, a normed IQ score with a 95% confidence interval. Not Mensa-qualifying, but a realistic estimate of your latent ability.
Take the Core Brain IQ test →