Your results
The overall score
One number out of 10 that estimates how a typical group of people would rate your photo. Here's exactly where it comes from.
What goes into it
Two AI vision models each rate seven features from 1 to 10. We average the two models feature by feature, so no single model decides your score.
| Feature | What it rates |
|---|---|
| Symmetry | How closely the left and right sides mirror each other |
| Canthal tilt | The angle of the eyes from inner to outer corner |
| Jawline | Definition and shape of the jaw |
| Cheekbones | Height and projection of the cheekbones |
| Skin | Clarity, texture and evenness |
| Eyes | Shape, openness and the area around them |
| Nose | Size and shape relative to the face |
From ratings to a score
The average of the seven feature ratings is then calibrated. AI models rate faces in a compressed range: they rarely give very low or very high marks. Calibration stretches their average so scores spread out the way human ratings do, with a typical face landing around 6.0.
The result is rounded to one decimal and kept between 1 and 10.
What each range means
| Score | Label | Where it sits |
|---|---|---|
| 4.0 | Slightly below average | Top 92% |
| 5.0 | Average | Top 76% |
| 6.0 | Above average | Top 50% |
| 7.0 | Clearly above average | Top 24% |
| 8.0 | Striking | Top 7.7% |
| 9.0 | Exceptional | Top 1.6% |
"Top X%" compares your score with the bell curve above: a 7.0, for example, scores higher than roughly three in four people.
Why your score can change between scans
- Lighting: soft front light flatters skin; harsh overhead light deepens shadows under the eyes.
- Lens distance: close-up selfies distort proportions (see the photo guide).
- Expression and angle: a slight smile or head tilt changes how features read.
- Model variation: AI ratings vary a little run to run, which averaging two models reduces.