An 8 out of 10 is rare. Across 17,032 AI face analyses, only 3.3% of people scored 8.0 or above, against a measured average of 5.59. In raw numbers, 524 people landed in the 8.0-8.9 band and just 36 reached 9.0 or higher.
It is also the level at which the question changes from "how do I gain points" to "how do I avoid losing them".
How rare is an 8 out of 10?
Rare enough that the percentiles start to bite. The 90th percentile of all scores is 6.8, the 95th is 7.8 and the 99th is 8.8 - so an 8.0 is already past 96 or 97 people in 100, and an 8.8 is genuinely top 1%.
The tail thins fast above that. Only 0.2% of the dataset, 36 people out of 17,032, scored 9.0 or better. For comparison, 4,230 people scored in the 4.0-4.9 band alone - meaning for every person who scores a 9, there are more than a hundred who score a 4.
What an 8 out of 10 actually measures
The most useful thing in the data is not how high an 8 scores, but how evenly. Every analysis scores twelve features separately, and comparing what 8+ scorers average on each against the population average shows the pattern.
| Feature | 8+ average | Everyone | Difference |
|---|---|---|---|
| Bone structure | 8.90 | 5.60 | +3.30 |
| Jawline | 8.89 | 5.59 | +3.30 |
| Facial harmony | 8.66 | 5.58 | +3.08 |
| Eyes | 8.62 | 5.97 | +2.65 |
| Symmetry | 8.54 | 5.62 | +2.92 |
| Chin | 8.48 | 5.39 | +3.09 |
| Skin | 8.33 | 5.30 | +3.03 |
| Eyebrows | 8.27 | 5.94 | +2.33 |
| Cheeks | 8.25 | 5.42 | +2.83 |
| Nose | 8.01 | 5.36 | +2.65 |
| Lips | 8.00 | 5.65 | +2.35 |
| Forehead | 7.66 | 5.39 | +2.27 |
The spread across an 8+ face is only 1.24 points, from 8.90 at the top to 7.66 at the bottom. There is no weak link. That is the actual structure of a high score: not one spectacular feature compensating for ordinary ones, but twelve features all sitting high at the same time.
It is the mirror image of what happens lower down the scale, where the average gap between a person's best and worst feature is 1.38 points and 58% of people tie for their weakest.
Why does nobody score a 10?
Because the things a 10 would require do not co-occur in real faces.
Facial symmetry is the strongest single correlate of the overall score, and only 3.3% of faces score 8 or above on symmetry alone - just 0.9% reach 9. Golden-ratio compliance averages 77.28% across the dataset and reaches 89.03% in the 8.0-8.9 band, but the single highest value recorded across all 17,032 faces is 96%. Nobody hits the "perfect" 1.618.
A perfect 10 would need every one of twelve measured features at the top of its range simultaneously, in the same face, in the same photograph, with symmetry and proportion both at values that essentially nobody reaches. It does not happen. A tool that hands out 10s is not measuring anything - it is telling you what you want to hear.
Is there anything left to improve at an 8?
Very little, and the data is blunt about it. The 8.0-8.9 band averages 0.39 points of headroom - the smallest of any band except 9 and above, which averages 0.32 - with an average potential of 8.84.
That is the mirror image of the low end, where the 4.0-4.9 band carries 1.75 points. High scorers have already converged on their ceiling: the skin is clear, the hair works, the body composition is handled, the photograph is competent. There is very little left on the table.
What remains is maintenance rather than improvement, and the risk profile inverts. At an 8, a bad night's sleep, harsh overhead lighting or an unflattering angle can each cost more than the 0.39 points that separate you from your own ceiling. The realistic goal is consistency - scoring your 8 reliably rather than occasionally.
The limits of a high score
The same caveats apply here as anywhere else on the scale, and they are worth stating precisely because a high score is the one people are most tempted to take as a verdict.
This is one model scoring facial geometry and presentation from a single photograph. It does not measure charisma, expression in motion, voice, style or how someone actually comes across, and those carry enormous weight in real attraction. An 8 in a photograph is not a guarantee of anything in a room.
It is also worth remembering that people who use a face-rating tool are not a random sample of the population. A 3.3% figure describes the people who showed up, not humanity.
The full percentile table is in the average face rating study, the symmetry and golden-ratio figures come from the facial symmetry study, the feature-level data is from the which facial feature matters most study, and the headroom figures are from the potential vs actual study. To see your own current and potential scores, run the face rating tool, or take the attractiveness test to get the same score with your percentile against the distribution above.