Every year or so a link goes around that promises to settle the camera argument: a page of unlabelled photo pairs, one tap each, and a ranking of the year's phones at the end with every logo stripped off. MKBHD's blind smartphone camera test is the closest thing this industry has to a popular vote, and its results get quoted for years afterwards as a verdict on hardware.
They are a verdict on something narrower, and the narrower thing is more useful to you than the headline. What follows is how the test is built, what the votes actually rank, and what any of it says about the phone in your pocket.
How the test works
The shape is the same every edition. A set of phones shoots the same scenes. The files are stripped of identification and assigned letters, the voting page puts two of them side by side, and you pick the one you prefer. Nobody voting learns which phone took either photo until the results come out.
Nothing is scored against a checklist. Each tap is a head-to-head result, and the phones are ranked with an Elo rating — the mechanism chess uses, where beating a strong opponent counts for more than beating a weak one. Rounds run in three categories: standard daylight, low light and portrait mode.
The scale is what makes it worth reading. For the 2023 edition, MKBHD wrote on his own account that 20 phones were tested and more than 20 million votes were collected. He posted the category winners in December from what he described then as a few million votes, and those podiums held when the full count landed in January. The results are in The Blind Smartphone Camera Test Winners! on his channel.
Nobody swept, and that is the finding
The 2023 podiums, as he published them:
- Standard: Pixel 7A, then Pixel Fold, then OnePlus Open.
- Low light: iPhone 15 Pro, then Pixel 8 Pro, then Pixel 7A.
- Portrait: Pixel 8 Pro, then Z Fold 5, then iPhone 15 Pro.
Nine places, six different phones, and no phone taking first twice. That is the most quotable thing about the exercise and almost never the thing that gets quoted. The winner changes with the category, which means the question "which phone has the best camera" does not have one answer even inside a single test with twenty million votes behind it.
The Elo scores behind that final ranking sit close together at the top of each category: tens of points between first and third, not hundreds. A podium is a ranking of near-ties presented as a hierarchy.
What a tap actually records
Here is what a voter is actually doing. Two web-compressed images appear at whatever size the screen gives them, often a phone, and the voter answers "which looks better" quickly — the only way to get through a round.
That question rewards a specific set of processing choices: a brighter exposure, more contrast, more saturation, a face lifted cleanly out of its background. It does not reward the things that decide whether you keep a photo — detail that survives a crop, colour that stays consistent across twenty frames of the same afternoon, shadows that hold up when you edit them later.
None of that makes the result wrong. It makes it specific. The test measures which processing choices win a fast, small, side-by-side look, across an enormous sample of ordinary eyes. That is worth knowing, and it is not the same as which camera is better.
What it says about the phone you already have
The useful part for you is the part that embarrassed the flagships. A mid-range phone won the standard category outright and finished on the low-light podium as well, ahead of devices several tiers above it. That pattern is not a fluke of one year, and it is worth understanding on its own terms.
If your phone is two or three years old, or was never the expensive one in the range, take the implication seriously: at the size and in the conditions this test presents, the gap to the current flagship is often invisible to a stranger hunting for it. Upgrading for a better photo of a friend outside a café is a weaker case than the marketing suggests.
Where the newer phone genuinely pulls away is somewhere the test never goes: zoom beyond the main lens, video, shutter lag, and how often it gets the shot right on the first try rather than the third.
What it cannot tell you
Four limits, all of them structural rather than sloppy. Each category rests on the same small set of scenes: every phone shoots the same photo, which is what makes the comparison fair and also what makes it narrow. Only stills are voted on. Processing is software, so a phone's placing describes the firmware it shot with, not the phone it is today after a year of updates. And your taste is not the median voter's: a test that aggregates millions of opinions is measuring the middle, and nobody actually stands there.
Read it as what it is: a vast preference survey on camera output, conducted honestly, with the logos off. Then look at your own last hundred photos, a better-matched sample of your life than any scene MKBHD could have chosen.
The interesting follow-up is the disagreement: when the crowd and the reviewers pick different winners, it is usually the question that differs, not the eyesight. And before you take any creator's numbers as a spec, what a big-audience test can and cannot establish is the piece that keeps results like these in proportion.
Last updated September 22, 2026