Jawline Rating: Measure Your Jaw, Chin and Cheeks From One Photo

FaceCheckup finds 478 landmarks on your face with Google's MediaPipe Face Landmarker, running on your device so the photo never leaves it. Three of the outputs describe the lower face. Jaw width ÷ cheekbone width compares the distance between the two jaw corners with the distance between the widest points of the cheekbones. Forehead width ÷ cheekbone width does the same at the temples. Jaw angle (front view) is the angle at each jaw corner between the line running up to the cheekbone and the line running down to the chin, averaged across both sides.

A smaller front-view angle means the outline turns more sharply at the corner, which is what people usually mean by a 'sharp' or 'square' jaw. A larger angle means the outline flows more gradually from cheek to chin, which reads as softer or rounder. A wider jaw relative to the cheekbones makes the lower face look broader; a narrower one makes it taper toward the chin.

Those two numbers, width and corner, are what a camera can honestly capture. Definition under the jaw and chin depth need a profile view or a 3D scan, so we leave them out rather than guess.

Forum threads often ask for a 'gonial angle' rating. The gonial angle is a bone angle: it sits between the back edge of the lower jaw (the ramus) and its lower border, and it is measured from the side on X-rays or CT scans. Bakan and colleagues (2022) measured it on cone-beam CT images of 235 people in Turkey aged 7 to 77. The right-side gonial angle averaged 125.25° ± 6.61° in women (124 people) and 121.64° ± 6.88° in men (111 people).

Those figures are useful context, but they cannot be compared with anything a selfie shows. A front photo flattens the jaw into a 2D outline, and soft tissue sits over the bone. The angle you see from the front is a different angle in a different plane. That is why our measurement is labelled 'jaw angle (front view)' and why we never convert it into a gonial estimate. Any site that gives you a gonial angle from a front-facing selfie is inventing a number.

Run the scan above, then open the Jaw group of results. Switch on the Jaw layer to see the angle drawn at both corners, and the Widths layer to see the forehead, cheekbone and jaw lines laid across your face. If a line lands somewhere odd, such as inside a beard or below the jaw because of a shadow, you will see it immediately, which a bare score would hide.

Every number comes with a ± range. The scanner measures your photo plus three re-framed copies of it, and the ± is half the spread between them. In our own test on one selfie, canthal tilt read 7.01°, 7.81°, 7.31° and 7.46° across the four copies, a spread of about ±0.4°. Jaw numbers wobble the same way. Treat a difference smaller than the ± as no difference at all.

These numbers are shown without a 'normal range' next to them. The published jaw studies used calipers or CT scans on bone points that MediaPipe does not locate in the same place, so lining your number up against them would be comparing different things. The jaw measures also don't feed the Balance Score, which uses only symmetry and Pallett, Link and Lee's (2010) two eye ratios.

There is no large, multi-population norm for jaw width relative to cheekbone width. The best published numbers we found come from Virdi, Wertheim and Naini (2019), who measured 72 Kenyan-African adults aged 18 to 30 with calipers. Dividing their published means gives about 0.80 for men and 0.74 for women. That arithmetic is ours, the sample is small, and caliper points sit on bone landmarks rather than the photo outline, so the ratio cannot be laid over your result.

Face width itself is better documented. In the US Army's ANSUR II survey of 6,068 adults, cheekbone (bizygomatic) breadth averaged 142.6 mm in men and 133.8 mm in women, calculated by us from the public data. Your phone cannot give you millimetres at all: a photo has no absolute scale, and the landmark model's own error is about 2% to 3.9% of the distance between the eyes, according to Google's model card. That is why every jaw figure we show is a ratio.

Looksmaxxing forums lean heavily on the facial width-to-height ratio (fWHR) as a 'masculinity' number. The research is far less dramatic. A meta-analysis by Geniole and colleagues (2015) found that fWHR was larger in men than in women by d = .11 across 10,853 people, which is a very small difference. Summersby and colleagues (2022) measured 17,607 Australian passport photos and found men's fWHR larger only in young adulthood, with women's larger after age 48.

Raw width tells a clearer story. Caton and Dixson (2022), working with the same 6,068 ANSUR II adults, found bizygomatic (cheekbone) width highly sexually dimorphic, with an equivalent effect size of d = 1.39. They also showed that ratio measures like fWHR can produce misleading results, because the sign of the sex difference flips depending on which height you divide by. The CT data above show men's gonial angles slightly smaller than women's on average, with wide overlap.

So when someone says a jaw looks 'masculine', they are mostly reacting to overall size and width, which a ratio from a selfie cannot capture. A ratio describes shape, not sex, testosterone or personality.

Stylists judge face shape by comparing three widths, forehead, cheekbones and jaw, plus face length and how sharp the jaw corner is. FaceCheckup's face-shape detector uses exactly those inputs, which is why your jaw result feeds into the shape you get. Round and square faces have about the same length and width in every stylist definition we collected; the one thing that separates them is the jaw corner. Heart and diamond faces both narrow toward the chin, and diamond adds a narrow forehead so the cheekbones stand out as the widest point.

If your jaw ratio is high and the corner is sharp, expect a square or triangle (pear) result; if the jaw is narrow, expect heart or diamond. The detector shows the top two shapes with percentages instead of forcing one. The high cheekbones, face shapes and male face shapes pages go deeper into each part of this outline.

Many jawline searches end in advice to mew, chew hard gum, or worse. A 2025 systematic review by Seraj and colleagues found a paucity of studies on mewing, with claims resting on anecdotal or low-level evidence. A 2026 study by Grillo and colleagues tracked 'bonesmashing' (hitting the face to provoke bone growth) on social media and described it as a dangerous trend being normalised with pseudoscientific language.

We don't give advice on exercises that claim to reshape bone, on fillers, on surgery or on any medical procedure. If you are thinking about a procedure, talk to a licensed medical professional. Our measurements are not medical: Google's model card for the landmark model states it is not intended for life-critical decisions. What we can offer instead are choices that are free, reversible and sourced: facial hair, haircuts, frames, makeup and the way a photo is taken.

A beard is the most direct way to change the outline of the lower face, and barbers agree on a few rules. Each one below is a stylist convention, sourced from at least two grooming publications (FashionBeans, The Art of Manliness, Philips, Pall Mall Barbers, Harry's and Braun).

Where hair ends matters more than the cut's name. Stylists at Revlon Professional, Matrix and InStyle agree that a wide jaw looks best when the hem or layers end below the jaw, not exactly on it, because a blunt line at the jaw angle draws attention to its width. A narrow jaw is the opposite case: fullness at chin level, such as a wavy chin-to-collarbone bob or a lob, fills out the lower face. For men with a square face, barbers suggest height and texture on top with neat sides (all stylist conventions).

Opticians use frames to move attention up or down. If your jaw is wide, WebMD, All About Vision and Vision Express suggest putting the frame's visual weight on top, such as a browline or a bold top rim. For a square face, curved oval or round frames contrast with the jaw's angles, while boxy, sharply angular frames echo them (optician conventions).

If you wear makeup, contour follows the same logic. For a square face, makeup artists at e.l.f., Maybelline and AARP shade the outer corners of the forehead and the jaw corners, with only a light touch under the cheekbones. For a round face, Patrick Ta, e.l.f. and AARP place contour along the jaw and in the hollows under the cheekbones, with highlight down the centre. For a heart face, keep contour off the jaw, which would narrow the lower face further, and shade only the tip of the chin (makeup-artist conventions).

Camera distance changes faces more than most people expect. Ward and colleagues (2018) modelled a selfie taken about 12 inches away and found the nose base appears about 30% wider relative to face width than in a photo from 5 feet. Bryan, Perona and Adolphs (2012) photographed the same faces at 45 cm and 135 cm; the close photos were rated less attractive, trustworthy and competent. Both effects change how the lower face reads, so measure and share photos taken from farther away.

Start with the measurement, then pick one change. Take a straight-on photo with the phone propped at eye level, hair pulled back, glasses off, mouth closed and a neutral face. Run the scan and switch on the Jaw and Widths layers. Note the jaw ratio, the front-view angle and their ± ranges, and check whether the face-shape result lands on square, round, heart or somewhere between two shapes.

If you have facial hair, try the neckline rule first: it costs nothing and changes the jaw outline within one trim. Booking a haircut? Tell your stylist where the hem should land relative to the jaw. Then retake the photo the same way and compare. The numbers will barely move, because bone doesn't change, but you will see how much of what people notice is outline, shadow and angle.