Average IQ: what it is and how to read the IQ table
Updated October 26, 2026 · QI Agora team
The average IQ is 100. That's not a coincidence or an average worked out every now and then: modern tests are built so that the comparison group's average is exactly 100, with a standard deviation of 15 points.
IQ table: ranges and share of people
The table below uses QI Agora's ranges and the approximate share of people in each range under a normal distribution (the famous "bell curve").
| IQ | Range | People in this range |
|---|---|---|
| 145 or more | Exceptional | about 0.1% |
| 130 to 144 | Well above average | about 2% |
| 115 to 129 | Above average | about 14% |
| 85 to 114 | Average | about 68% |
| 70 to 84 | Below average | about 14% |
| below 70 | Well below average | about 2% |
Range names vary from test to test. They are a simple way to communicate the result, not a label for anyone.
Percentile: the easiest way to read your result
The percentile tells you how many people out of 100 scored lower than you. Some examples on the scale with an average of 100 and a deviation of 15:
- IQ 100 → 50th percentile (half of people below);
- IQ 110 → about the 75th percentile;
- IQ 115 → about the 84th percentile;
- IQ 120 → about the 91st percentile;
- IQ 130 → about the 98th percentile.
On QI Agora, your result shows the estimated IQ, the range and this percentile.
"What's the average IQ by country?"
It's a common question, but the honest answer is: it depends on the test and the comparison group. The "IQ by country" rankings that circulate online mix different tests, small samples and distant years, and are heavily criticized by researchers. In a well-calibrated test, the reference group's average is always 100.
How QI Agora gets to the number
The test uses Item Response Theory, the same family of methods used in large-scale exams: each question has a difficulty, and getting a hard one right counts more than an easy one. Response time weighs a little, only on correct answers. At first, IQ is calculated against a theoretical reference; as more people take the test, the calibration switches to the site's real data. See the details.