Every other ranking on this site asks what a thing does. This one asks what it looks like on a desk you sit
at, and then reports the bill.
The design axis is not a taste score. It reads build quality and how an object sits in a room together —
whether a case flexes, whether a shape was resolved from every angle or only from the top, whether the
thing still looks like itself after a year of hands on it. Sorting by that axis alone produces an order
with no opinion whatsoever about comfort, and the gap that opens up is the point of the page.
Read every entry with its second number attached. It is in each note, and in two places it is brutal.
The gap this page exists to show
The Magic Mouse takes second place on looks and sits in the bottom handful of the entire tech category on
ease of use. Nothing else on this site splits that far.
Second place is earned. It is the only entry here resolved from every angle rather than from the top,
and making the entire upper face a touch surface removes the wheel, the seam and the button split that
every rival on this page still has to arrange around. On an axis that reads build and presence together,
very little is left to mark it down for.
The other number comes out of the same decision, which is the part that matters for a design-sorted
list. Flatness is what lets it sit almost level with the desk, and flatness is what the forearm pays
for; the product page sets out the posture and how long it takes to
register. This is not a manufacturing lapse waiting for a revision to fix. It is one shape, scored twice,
and the two scores were always going to point in opposite directions.
The charging port underneath is the smaller version of that and the harder one to defend, because it
survived the move to USB-C. A constraint gets accepted once. This one keeps getting chosen.
Materials decide the top, photographs decide nothing
First place is the heaviest thing on this page by a wide margin. A CNC aluminum case puts the Keychron Q1
close to two kilograms, and the consequence is that the board does not move, does not flex, and does not
sound hollow when you bottom out. Gasket mounting suspends the plate in silicone instead of screwing it to
the case, and the stabilizers arrive lubricated. Those are not visible in any product shot, and they are
most of why it wins.
The Keychron K2 sits just below the cut, and the comparison is instructive. Same layout, same brand, same
general silhouette, and a plastic case that flexes slightly under heavy typing with a hollower sound to
match. In photographs the two boards are near-identical. In the hand they are not close, and the design
axis is scored from the hand.
That is the working rule for this whole category. Anodising, RGB, and a sculpted silhouette are cheap.
Rigidity, a case that stays put, and a sound that is not hollow are not.
Small is not the same as restrained
The MX Anywhere 3S is the most finely made object on this page. It carries the electromagnetic scroll
wheel from the mouse two places above it, the same quiet switches, and a sensor that tracks on a glass
table where ordinary optical mice produce nothing at all. As a piece of engineering it outclasses most of
what is above it.
It ranks fifth because it was built to survive a bag, and a shell built for a bag is small. Across a full
day the hand curls tighter around it than it would around a full-size shape, which is exactly what its
ease-of-use score records. On a minimal desk it photographs as restraint. By the evening it reads as
cramped.
The side buttons are the smaller version of the same story: close together, easy to catch both at once
with a larger thumb, and a consequence of the size rather than a defect. It is the entry most likely to
look right in the shop and disappoint in month two — a different failure from the Magic Mouse, which
disappoints on day one and gets forgiven anyway.
Part of the design is not on the desk
Two entries here need a resident application to be the thing they are sold as.
The Apex Pro TKL earns its third place honestly on hardware: an aluminum top plate, double-shot PBT
keycaps, per-key lubrication, and layered foam that puts its acoustics near the top of prebuilt boards.
The small OLED handles presets and media without opening anything. But the per-key actuation depth, the
feature the board exists for, is configured in the SteelSeries suite, and that suite runs in the
background. The MX line has its own version of this in Options+, which owns gestures and per-application
mapping.
Score that the way this site scores any driver dependency. The hardware will outlive the support window.
When it closes, the case is still beautiful and the interesting half of the product is gone. That is why
the top of the durability-sorted keyboard page looks
different from the top of this one, and why a board with no software at all finishes second over there
while not appearing here.
The two that ask for nothing
Fourth and sixth place are the useful entries for anyone who came to this page wanting to buy on looks and
would rather not pay for it.
The MX Master 3S was sculpted rather than styled. The high back and long body exist because a palm grip
needs support, and the result reads as deliberate on a desk while scoring 8.2 on ease of use — the highest
of the top four. The cost is real but it is a fit problem rather than a comfort problem: small hands, claw
grips, and left-handed users are all out, and no software setting recovers a shape mismatch.
The Lift finishes last here and first on the axis this page ignores. Its 57-degree angle rotates the
forearm toward a handshake position, which looks strange on a desk for the first fortnight and then stops
registering as strange at all. Expect one to two weeks before precision comes back to normal. There is a
wrist-strain page where it does considerably better than sixth.
How to read this order
Nothing here is ranked wrongly. The axis was named before the list was written, and the list follows it.
What the order cannot tell you is duration. Two of the top three are objects that reward looking, and both
charge for it — one immediately, one after about three hours. The two at the bottom of this page would
finish higher on any ranking that counted the afternoon, and one of them
already does.
So use this page to narrow, not to decide. If two candidates are close on everything else, the one that
looks right is the one you will keep on the desk. If a candidate wins here and loses everywhere else, you
have not found a good product. You have found a good-looking one, which is a smaller thing.