AI Facial Analysis: Celebrity Face Scores — Who's a 10/10?

In our AI facial analysis, a 10/10 (a 95-100 overall) is rare — the math is unforgiving. The closest among the 52 celebrity face scores on this page sit in the low-to-mid 90s, led by the top of the leaderboard above. The AI grades geometric proportions (symmetry, golden ratio, canthal tilt, FWHR and 13 more metrics), not fame or charisma, so most A-listers land in the 80s rather than a perfect 10/10.

AI facial analysis maps 68 landmarks on a face and measures 17 objective proportions — facial symmetry, golden ratio adherence, canthal tilt, facial width-to-height ratio, jawline angle and more. The weighted average becomes a 0-100 celebrity face score: a geometry-conformance number (how close the measured proportions sit to ideal ranges), not an attractiveness ranking. Every score on this page is an AI estimate from a public reference photo, not a statement of fact about the person.

"Face IQ" (often typed "FaceIQ") is a generic term people use for an attractiveness or face-rating score. The celebrity scores on this page are produced the same way — 17 facial-geometry metrics (symmetry, golden ratio, canthal tilt, FWHR and more) condensed into one 0-100 geometry-conformance number (proximity to ideal ranges, not an attractiveness verdict) — so if you searched "Face IQ celebs," this leaderboard is the celebrity face-IQ ranking you were after. Every score is an independent AI estimate from a public reference photo, not an official rating from any other app or brand.

Henry Cavill's face rating here is 94/100 — the highest of the 52 celebrity faces analyzed on this page, with jawline angle his strongest single metric at 97. Like every score on this leaderboard, it is a geometry-conformance number from 17 facial measurements (symmetry, canthal tilt, FWHR and more) — an AI estimate from a public reference photo, not an attractiveness verdict about the person.

We run the same client-side facial analysis used in our looksmaxxing test on a public photo of each celebrity. The AI maps 68 facial landmarks and computes 17 objective metrics — facial symmetry, canthal tilt, FWHR, golden ratio, jawline angle, hunter eye index, and more. The overall score is a weighted average of these metrics measured against published reference ranges.

The AI measures geometric proportions, not perceived attractiveness or fame. Henry Cavill scores high because his facial symmetry, canthal tilt, FWHR, and golden-ratio alignment all sit close to the geometric ideal. The Weeknd has a less symmetrical structure with a softer jawline angle by the math — even though many people find him attractive subjectively. Math ≠ taste.

The geometric measurements are precise — the same algorithm gives the same score on the same photo every time. What the score does NOT measure is subjective attractiveness, charisma, style, voice, or personality. Use the celebrity scores as a benchmark for what specific facial-metric combinations look like, not as a verdict on who is "more attractive."

Facial symmetry, canthal tilt (eye angle), facial width-to-height ratio (FWHR), golden ratio alignment, facial thirds balance, eye spacing, jawline angle, midface ratio, nose proportion, lip ratio, hunter eye index, brow-to-eye distance, and 5 more secondary metrics. All are standard biometric measurements used in cosmetic surgery and beauty research.

Yes — our free looksmaxxing test runs the exact same 17-metric engine on your photo. Upload a front-facing photo and you'll get the same per-metric measurement map, plus something the celebrity leaderboard doesn't have: a validated Face Score percentile (cross-validated at r ≈ 0.8 against 5,500 human-rated faces). On desktop the scan runs in your browser; on mobile your photo is processed in memory by our scan server and deleted immediately after — never stored.

Photo conditions matter. Studio lighting, angle, makeup, and facial expression can change the measured canthal tilt and symmetry by ±5 points. We use a high-resolution, front-facing, neutral-expression photo for each celebrity to keep comparisons fair. The same celebrity in a candid photo would score slightly differently.

No. QOVES uses a paid human-led analysis ($150/year) with subjective elements. Photofeeler uses crowdsourced human votes. RealSmile uses fully objective AI geometry — no humans, no subjectivity, no waiting. The trade-off: we measure structure, not perception.

We rank 52 frequently-discussed celebrities. If a face you expected isn't here, it's either on our extended list (running for new releases monthly) or simply hasn't been added yet. Score-wise, very few public figures actually break 95 — the math is unforgiving.