Face Shape Rarity Ranked: What 51,000 AI Scans Reveal
Oval is the most common. Diamond is the rarest. And the shape closest to the golden ratio isn’t the one you’d guess. Here’s the full breakdown, shape by shape.
Key Takeaways
- Oval is the most common face shape at 28.4% of the dataset — but it’s far from a majority.
- Diamond is the rarest, at roughly 5%, followed by triangle at 8.6%.
- Oblong scores closest to the golden ratio (84/100 average), not oval — a gap between what’s called “balanced” and what’s mathematically closest to phi.
- Round face shapes split sharply by gender — 22% of women vs. 12% of men.
- Check your own shape against this dataset with the free AI face shape detector.
Contents
Methodology: How the Data Was Collected
The dataset comes from OblongFaceShape.com’s free AI face shape detector, which classifies face shape via photo upload, live camera, or manual measurement entry. Every one of the 51,247 analyses to date contributes an anonymous data point.
A Google MediaPipe FaceLandmarker model maps 478 facial landmark points directly in the browser, calculating four proportions — forehead width, cheekbone width, jawline width, and face length — the same landmark system used in craniofacial anthropometry (Farkas, 1994). A published, transparent formula then classifies each face into one of seven categories. Photos are processed on-device and never stored — the dataset holds only the resulting measurements, not images. Full methodology: face shape statistics page.
Oval Face Shape
28.4% · Most CommonRatio band: 1.25–1.45 · Cheekbones widest, gently tapered jaw
Oval is the shape most often called the cosmetology “reference” proportion, and the data backs that up — it’s genuinely the most frequent classification. But at just over a quarter of all faces, it’s nowhere near a majority, which is worth knowing before assuming oval is the default.
Take the free oval face shape test →Oblong Face Shape
21.7% · 2nd Most CommonRatio band: 1.5+ · Even widths top to bottom, straight sides
Oblong is the second most frequent shape in the dataset and, notably, the one that scores closest to the golden ratio (more on that below). It gets comparatively little dedicated styling coverage relative to how common it actually is.
Read the full oblong face shape guide →Round Face Shape
17.3% · 3rd Most CommonRatio band: 1.0–1.15 · Soft contours, no jaw angle
Round shows the sharpest gender split of any category: 22% of women vs. 12% of men — nearly double. That tracks with anthropometric research on male jaw and brow development reducing “soft contour” classifications.
Take the free round face shape test →Heart Face Shape
14.1%Forehead 15%+ wider than jaw · Often a widow’s peak
Fourth most common overall — common enough for dedicated styling coverage, but proportionally distinct enough (via the forehead-to-jaw ratio) that it’s rarely confused with oval or round in the raw measurements.
Take the free heart face shape test →Square Face Shape
9.5%Jaw width within 10% of forehead · Angular mandible
Less common than its prominence in styling content suggests. “Square” gets used loosely as shorthand for any strong jaw, but the strict definition — near-equal forehead and jaw width plus visible angularity — is a narrower band than the casual usage implies.
Take the free square face shape test →Triangle Face Shape
8.6% · 2nd RarestJaw averages 18% wider than forehead · Base-heavy silhouette
The inverse of heart-shaped — jaw-dominant rather than forehead-dominant. It gets noticeably less mainstream styling coverage than its close cousin, square, despite representing a meaningfully different proportional profile.
Read the full triangle face shape guide →Diamond Face Shape
~5% · RarestCheekbones average 22% wider than jaw · Most angular of all 7
The rarest of the seven shapes by a clear margin, which likely explains why diamond-specific advice (particular frame shapes, particular contouring) tends to be more narrowly targeted than advice for higher-frequency shapes.
Take the free diamond face shape test →Which Face Shape Is Closest to the Golden Ratio?
The golden ratio (phi, ≈1.618) is often cited in facial aesthetics literature as a marker of proportional harmony, though its link to perceived attractiveness remains debated (Pallett et al., 2010). In this dataset, oblong — not oval — scores highest on golden ratio proximity. That’s a real gap between the shape most often called “balanced” in consumer content and the shape that’s numerically closest to phi. Score your own proportions with the golden ratio face calculator.
Do Face Shapes Differ Between Men and Women?
The clearest split in the data: round face shapes at 22% of female analyses vs. 12% of male — nearly double. This lines up with craniofacial anthropometry research showing male facial development tends to produce more pronounced jaw angularity and brow prominence on average (Enlow & Hans, 1996), which reduces “round” classifications — a category defined by the absence of jaw angularity — among men. No equivalent split was published for the other six categories, which is worth flagging rather than glossing over.
Limitations of the Data
This is a self-selected consumer sample, not a randomized population study. People who seek out a face shape detector online likely skew toward users actively researching hairstyles, glasses, or beauty content — not a random cross-section of the general population.
The classification method is transparent and reproducible, but it’s a proportion-ratio formula, not a peer-reviewed clinical diagnostic. Photo-based accuracy is reported at 85–92%, vs. 88–94% for direct measurement entry — so individual classifications carry some uncertainty even where the aggregate distribution is likely directionally reliable at this sample size.
Frequently Asked Questions
What is the most common face shape?
What is the rarest face shape?
Which face shape is closest to the golden ratio?
Does face shape differ between men and women?
How was the face shape data collected?
Can I find my own face shape using the same method?
Rizwan Aslam
Founder of OblongFaceShape.com and developer of the site’s face shape analysis methodology, informed by peer-reviewed facial anthropometry research and used by over 51,000 people worldwide.
About the author →References
- Farkas, L.G. (1994). Anthropometry of the Head and Face (2nd ed.). Raven Press.
- Enlow, D.H. & Hans, M.G. (1996). Essentials of Facial Growth. W.B. Saunders Company.
- Pallett, P.M., Link, S., & Lee, K. (2010). New “golden” ratios for facial beauty. Vision Research, 50(2), 149–154.
- OblongFaceShape.com (2026). Face Shape Statistics: 51,247 AI Face Scans. oblongfaceshape.com/face-shape-statistics
