How close your face sits to 1.618, and the two proportions research found actually rated best. Those are near 36% and 46%, they have nothing to do with phi, and all three are reported here.
Choose one clear, front-on photo
Looking at the camera, even light, hair off your forehead, no
glasses.
JPG, PNG or WebP up to 10 MB, or drag one here
Your photo never leaves your device.
Proximity to phi
%
Add a photo and this fills in.
Far from phi below 40
Some way off 40 to 70
Close 70 to 85
Very close 85 to 100
Eye-to-mouth, as a share of face length
not yet %
Eye spacing, as a share of face width
not yet %
Face length to width
not yet
Typically 100% would be an exact match to phi
Taking the measurement
1
Take a front-on photo
Square to the camera, eyes level, even light, no glasses.
2
Choose the file
Read in your browser. Nothing is uploaded.
3
Compare the two answers
The phi figure and the Pallett figures measure different things. Where they disagree, the second pair is the one with experiments behind it.
What phi claims
Example
Face length against face width. The claim is that this lands near 1.618.
What the research found
Example
Pupil to pupil across the face, and the eye line down to the mouth.
Drawn from the landmarks our own measurement reads off this photo, not
placed by hand. It is a generated face, not a real person, and nothing
here is a verdict about it.
This measures your face length against its width and reports how close that sits to 1.618. It also measures the two proportions researchers found actually rated best, which are about 36% and 46% and have nothing to do with phi. Both numbers are shown because only one of them has evidence behind it.
How this is worked out
Proximity to phi
Face length divided by face width, compared with 1.618. This is the figure the term "golden ratio face" refers to, and it is here because you searched for it.
Eye-to-mouth, as a share of face length
Pallett and colleagues found faces rated best when this sat near 36%. This is a measured optimum from four experiments, unlike the phi figure above.
Eye spacing, as a share of face width
The same study found about 46% rated best. Both proportions turned out to match those of an average face.
What each measurement is
Reference bands are this tool's own, chosen for the landmarks it uses.
Where research reports a figure it is given separately, with the source,
because the two are different claims. And each one says what actually
moves it, which for most facial measurements is the more useful half.
Proximity to phi %
Formula
how close face length ÷ face width sits to 1.618
Reference
100% would be an exact match to phi
Research
a systematic review found the golden proportion does not consistently predict facial attractiveness across populations - Londono et al. (2024), Journal of Prosthetic Dentistry
Measured because you searched for it. The two rows below are the ones with experiments behind them.
What moves it Nothing to work toward. The review found it does not predict attractiveness, and the proportion is skeletal in any case. The two rows below are where the real figures are.
Eye-to-mouth, as a share of face length %
Formula
pupil line to mouth line ÷ face length
Reference
36% on this tool's anchors
Research
about 36% rated best across four experiments - Pallett, Link & Lee (2010)
A measured optimum, unlike the phi figure above. It is also the proportion an average face has.
What moves it 36% is the closest thing to a real target in this whole section: it was measured, across four experiments, and it beat every alternative spacing tested. It is also fixed - the distance from your eyes to your mouth does not change. Worth knowing because it tells you what the research actually converged on, and because it is not 1.618.
Eye spacing, as a share of face width %
Formula
pupil-to-pupil distance ÷ cheekbone width
Reference
46% on this tool's anchors
Research
about 46% rated best - Pallett, Link & Lee (2010)
What moves it Measured, real, and fixed. Eye spacing is skeletal and does not change.
Face length to width
Formula
face length (top of face mesh to chin) ÷ cheekbone width
Reference
1.3 to 1.5 on this tool's anchors
Not the clinical facial index: the mesh finds the top of the forehead, not the hairline.
This measures how close your face sits to 1.618, because that is what “golden
ratio face” means and that is what you came for.
It also measures two other proportions. Those are the ones with experiments
behind them, and they are not phi.
The short version
The golden ratio is a real number with a long history in mathematics and art.
Its application to facial beauty is a twentieth-century marketing idea that
spread because it sounds rigorous.
A systematic review in the Journal of Prosthetic Dentistry settles most of it.
It found the golden proportion does not consistently predict facial
attractiveness across populations, and that it was not even prevalent in the
populations measured: three of the eight ratios examined departed significantly
from it.
That review is why the attractiveness test gives phi a
weight of zero in its formula while every competing tool weights it. Mask-based
scoring, the overlay approach, sits in the same category.
What researchers found when they went looking
The useful part is that somebody actually ran the experiment.
In 2010, Pallett, Link and Lee published four experiments in Vision Research.
They showed people faces with identical features and different spacings, and
asked which was more attractive. They were testing whether an ideal arrangement
exists at all.
It does. Faces were rated best when the vertical distance from eyes to mouth
was about 36% of face length, and the horizontal distance between the eyes
about 46% of face width.
Then they noted what those numbers correspond to. They are the proportions of
an average face.
So there is a measurable optimum in facial spacing. It is not 1.618. And it
lines up with a much older finding, Langlois and Roggman in 1990, who composited
many faces together and found the composites were rated more attractive than
almost all of the individuals that went into them.
The tool reports both your phi proximity and your two Pallett proportions, so
you can see the gap between the claim and the evidence on your own face.
Read the limits, because they are real
One study, female faces, spacing only. Pallett tested the arrangement of
features, not their shape. It is a single paper and we grade it as one.
Face “length” is ambiguous and that is why calculators disagree. Some
measure to the hairline, some to the top of the forehead. Those are far apart on
a receded hairline and they give different ratios. This tool measures to the top
of the detected face outline, which is the forehead rather than the hairline,
because a hairline is hair and no landmark model can find it.
Camera angle moves everything. A raised chin shortens the face in frame. A
phone held off-square compresses one side.
What this does not mean
A face close to 1.618 has not passed anything, and a face far from it has not
failed. The ratio is not a threshold and the research does not support treating
it as one.
The more interesting reading of the Pallett result is that the target everyone
is chasing turns out to be the ordinary case. Averageness is attractive, it
replicates, and it is the opposite of the rare-ideal framing that the golden
ratio is usually sold with.
Where to go next
The facial thirds tool covers the other classical canon
and the 1985 study that measured how rarely it holds. The
facial ratios calculator reports a dozen proportions at
once if you want the full set, and the
attractiveness test shows how these components combine
when they are weighted by evidence rather than by tradition.
Discussion of what any of this is worth is better had in the community than with
a calculator.
It is the idea that the most attractive faces have proportions close to 1.618, the number known as phi. It is repeated constantly in aesthetics marketing and it is poorly supported in the research literature.
Is the golden ratio actually linked to facial beauty?
The evidence is weak. What does have evidence is a different pair of proportions: Pallett and colleagues ran four experiments and found faces rated best when eye-to-mouth distance was about 36% of face length and eye spacing about 46% of face width.
Why does this tool measure phi at all if it is unsupported?
Because you searched for it, and a page that refuses to answer the question just sends you to a worse page. We measure it, show it, and then show you what the research found instead.
What are the real ideal facial proportions?
About 36% and 46%, from Pallett, Link and Lee. The striking part is what those correspond to: the proportions an average face already has, which matches the averageness research going back to 1990.
Is a golden ratio face mask accurate?
Mask-based facial scoring, the overlay approach, is in the same category as the ratio itself: widely repeated, poorly supported. Our evidence page lists both under what the research does not support.
What is the perfect face ratio?
The closest thing to an answer in the research is Pallett's: faces rated best with eye-to-mouth distance about 36% of face length and pupils about 46% of face width. Phi does not appear anywhere in that result, and a systematic review in the Journal of Prosthetic Dentistry found the golden proportion does not consistently predict facial attractiveness across populations.
Who came up with the golden ratio for faces?
The mathematical ratio is ancient. Applying it to facial beauty is a much later idea, popularised in cosmetic surgery marketing during the twentieth century, and it spread because it sounds rigorous rather than because it was tested.
Does my face need to match 1.618?
No. Faces vary enormously and the research does not identify a single correct proportion. The Pallett result is the closest thing to a measured optimum, and what it points at is the average rather than a rare ideal.
Is this the same as facial thirds?
Related but different. Thirds divide your face into three vertical zones and come from the classical canons. Our facial thirds tool measures those and explains how rarely they actually come out equal in real faces.
Is my photo uploaded?
No. Everything runs in your browser, and the site's content security policy blocks connections to any other domain.
Why do different golden ratio calculators disagree?
Because they measure different landmarks and none of them says which. Face "length" can run to the hairline or to the top of the forehead, and those give very different answers. Ours measures to the top of the face outline and says so.
Measure how your face divides into upper, middle and lower thirds, and see the 1985 study that found equal thirds are the exception rather than the rule.
