Yes. A 6 out of 10 is above average, and by more than most people assume. The mean face rating across 17,032 AI face analyses is 5.59, the median is 5.3, and only 25.1% of people score 6.0 or above - so a 6 puts you in roughly the top quarter of everyone who has taken the test.
It is also the last band where meaningful improvement is still routine. Above it, the numbers get unforgiving.
Why does a 6 feel worse than it is?
Because of an intuition that does not survive contact with the data. People treat 10 as the scale and therefore 5 as the midpoint, so a 6 reads as "barely scraping past halfway".
Real scores do not work that way. The median is 5.3 and 49.6% of everyone lands in the 5.0-5.9 band alone. The distribution is a tight cluster around the middle with thin tails, so every tenth of a point above the median moves you past a surprising number of people. A 6.0 is not "just over half" - it is the top 25%. The same score in Gen-Z terms is a face card that holds up rather than one that never declines, and the face card rating test reports it in that language against the same distribution.
How rare is a 6 out of 10?
2,891 people scored in the 6.0-6.9 band specifically, which is 17.0% of the dataset. Above that it thins out quickly: 4.8% land in 7.0-7.9, 3.1% in 8.0-8.9, and 0.2% at 9.0 or higher.
For calibration against the percentiles, the 90th percentile of all scores is 6.8. So a high 6 is already brushing the top 10%, and the difference between a 6.1 and a 6.8 is a much larger slice of the population than the 0.7 suggests.
A useful way to hold it: a 6 is common enough that you will meet plenty of people with one, and uncommon enough that three quarters of people do not have one.
What separates a 6 from a 7?
This is the most interesting question at this level, and the golden-ratio data answers it unusually clearly. Compliance measures how closely facial proportions track the classical ratios - it is a structural measure, not a presentational one.
| Score band | People | Golden-ratio compliance |
|---|---|---|
| 4.0 - 4.9 | 4,230 | 74.98% |
| 5.0 - 5.9 | 8,450 | 77.00% |
| 6.0 - 6.9 | 2,891 | 78.07% |
| 7.0 - 7.9 | 824 | 82.28% |
| 8.0 - 8.9 | 524 | 89.03% |
Look at the step sizes rather than the values. Moving from the 4 band to the 6 band - covering two whole bands and three quarters of the population - raises compliance by 3.09 points in total. Moving from the 6 band to the 7 band alone raises it by 4.21.
That single jump is larger than the entire climb beneath it. The reading is that below a 7, differences in score are substantially about presentation: skin, grooming, hair, body composition, photography. At the 6-to-7 boundary, measurable facial proportion starts doing much more of the work.
This is the honest reason the step from 6 to 7 is harder than the step from 5 to 6, even though both look like one point on the same scale.
Can a 6 become a 7?
Often, yes - but with the caveat above attached. The 6.0-6.9 band averages 0.87 points of headroom with an average potential of 7.53, and 7.0 is the threshold where you enter the top 8.1% of all scores. On the numbers, the average person in this band has a realistic path across that line.
The headroom is smaller here than lower down the scale, which is expected. Someone scoring a 6 has usually already got the basics right, so the remaining points sit in details rather than fundamentals: skin consistency, hair that suits the face shape, body composition, and photographing yourself at eye level in even light rather than under a ceiling bulb.
For comparison, the 4.0-4.9 band carries 1.75 points of headroom and the 8.0-8.9 band carries 0.39. A 6 is on the downslope of that curve - still real room, but less of it each step up.
Where the remaining points usually are
Skin, first. It is the weakest feature for 38% of people across the whole dataset, more than twice as common as the next-most-frequent weak point, and the lowest-scoring of the twelve features on average at 5.30. It is also fully improvable, unlike bone structure, chin or nose.
Eyebrows are worth naming too: they are the single most common strongest feature, best for 29.3% of people, and also fully improvable - which makes them one of the few places where a small change reliably moves a real number.
What a 6 will not become
A 9. The correlation between current and potential score across the dataset is 0.94, and only 15 people out of 17,032 showed a gap of 3 points or more. Nobody at all had a gap of 4.
It is also worth remembering what the score is and is not. This is one model reading facial geometry and presentation from a still photograph, with no access to how you move, speak or carry yourself. A 6 that photographs badly and a 6 that photographs well are the same face on this scale and very different faces in person.
The full percentile table is in the average face rating study, the headroom figures come from the potential vs actual study, and the compliance data is from the facial symmetry study. Run the face rating tool to see both of your own numbers, take the attractiveness test if you want the percentile alongside them, or read is a 7 out of 10 a good face rating for what the next band up actually requires.