The 1-10 Attractiveness Scale Explained (With Real AI Data From 38,000 Faces)
Everyone uses the 1-10 scale, but nobody agrees on what the numbers actually mean. We analyzed 38,000 faces with AI to find out what each score really looks like — backed by measurable facial metrics, not opinions.
TL;DR
Most people (68%) score between 4 and 6. The distribution is roughly bell-shaped.
A "7" is top 15-25% — genuinely above average, not "average" as the internet pretends.
True 9s and 10s are extremely rare — under 1% combined in our dataset.
Want to know your real score? Take the free looksmaxxing test — 17 metrics, percentile rankings, private.
Why the 1-10 scale is broken (and how to fix it)
The biggest problem with the 1-10 scale is that people use it inconsistently. On Reddit, a "7" somehow means average. On dating apps, everyone thinks they're an 8. In reality, most self-ratings are inflated by about 2 points compared to objective measurements.
To anchor the scale to something real, we ran our 17-metric facial analysis on 38,000 selfies and mapped the composite scores to a 1-10 distribution. The result is a data-driven scale where each number corresponds to a specific percentile range and a measurable set of facial characteristics.
What each number actually means (based on 38K faces)
1-2: Bottom 5%
Significant facial asymmetry, very atypical proportions across most metrics. Extremely rare in our dataset — most faces score above this range. These scores typically indicate measurement issues (bad photo angle, obstructions) rather than true facial characteristics.
3: Bottom 10-15%
Below average on most measured metrics. Noticeable asymmetry, non-ideal canthal tilt, or disproportionate facial thirds. About 8% of our 38K sample falls here.
4: 15th-30th percentile
Slightly below average. Decent symmetry but weaker scores on 2-3 key metrics like FWHR, jawline definition, or midface ratio. About 15% of faces land here.
5: 30th-50th percentile (true average)
This is genuinely average — not an insult, just the statistical middle. Reasonable symmetry, middling proportions. About 20% of our dataset. Most people who think they're a "7" are actually here.
6: 50th-70th percentile
Above average. Good symmetry, at least one standout metric (strong jawline, positive canthal tilt, or balanced facial thirds). About 20% of faces. This is where grooming and fitness start making a visible difference.
7: 70th-85th percentile
Genuinely attractive by the numbers. Strong symmetry, good FWHR, well-proportioned facial thirds. Multiple metrics in the above-average range. About 15% of faces. This is where most people want to be — and it's achievable with the right grooming and softmaxxing.
8: 85th-95th percentile
Top 10-15%. Near-ideal proportions on most metrics. Strong jawline, positive canthal tilt, near-perfect symmetry. About 8% of our dataset. These faces consistently score high across multiple independent metrics.
9: 95th-99th percentile
Exceptional. Near-ideal scores on 14+ of 17 metrics. Golden ratio proportions, perfect symmetry, optimal canthal tilt. About 3% of faces. Most professional models and actors fall in this range.
10: Top 0.1%
Statistically near-perfect across all 17 metrics simultaneously. In our 38,000-face dataset, fewer than 40 faces scored here. A true 10 is not just "very attractive" — it means near-mathematical perfection in facial geometry.
The distribution is a bell curve, not a flat line
One of the most common mistakes people make is treating the 1-10 scale as if scores are evenly distributed. They're not. Like height, weight, and IQ, facial metric scores follow a roughly normal distribution. The vast majority of people cluster in the 4-6 range, with progressively fewer people at the extremes.
From our 38,000-face analysis: about 5% score 1-2, 8% score 3, 15% score 4, 20% score 5, 20% score 6, 15% score 7, 8% score 8, 3% score 9, and under 1% score 10. If you've been told you're a "solid 7," you're probably being complimented more than you realize — that's top 15-25%.
What actually drives your score
A single number is less useful than knowing which specific metrics push your score up or down. The 17 metrics RealSmile measures include: bilateral symmetry, canthal tilt, facial width-to-height ratio (FWHR), jawline angle, midface ratio, facial thirds balance, eye spacing ratio, nose-to-face ratio, lip fullness ratio, chin projection, brow ridge prominence, and golden ratio adherence.
Some of these are fixed (bone structure), but many are improvable through grooming, posture, skincare, and fitness. A person scoring a 5 who improves their skin texture, loses face fat, and fixes forward head posture can realistically move to a 6 or 7 — that's a jump from average to top 30%. The looksmaxxing test shows you exactly which metrics to focus on.
Why a single number is not enough
Asking "what am I on a 1-10 scale" is like asking "how healthy am I on a 1-10 scale." The number alone tells you almost nothing. Two people can both be a "6" for completely different reasons — one has perfect symmetry but weak jawline definition, another has a strong jaw but asymmetric features.
That's why RealSmile breaks your score into 17 individual metrics with percentile rankings. Instead of "you're a 6," you get "your symmetry is 82nd percentile, your canthal tilt is 45th percentile, your FWHR is 71st percentile" — actionable data you can actually use.
Free · Private · Instant
Find out where you actually fall on the scale
17 facial metrics. Percentile rankings from 38K faces. Photos never leave your device.
Take the free looksmaxxing test →Keep perspective
The 1-10 scale measures facial geometry, not human value. Charisma, humor, confidence, kindness, and intelligence are invisible to any camera or algorithm. Many of the most attractive people in real life are not the ones with the highest facial metric scores.
Use the data to optimize what you can — your grooming, skincare, posture, and photo angles — and spend the rest of your energy on the things that actually matter in relationships and life.
Get your detailed facial metric breakdown
See where you actually fall on the scale — with data, not guesses. Free.