Research Explainer · XiaoHu Explains

GPT-5.6 Sol takes the top spot in web design rankings: it learned design taste — and how to dodge the usual AI clichés

Third-party benchmark Design Arena dug through a thousand of its web pages and found a gap in its design space, sitting exactly where purple gradients and other AI tells should be.
60-second briefing
  • On Design Arena's web design (non-agentic) leaderboard, OpenAI's GPT-5.6 Sol landed in first place, 18 spots above its predecessor GPT-5.5 — the first time an OpenAI model has topped this board.
  • The benchmark team converted 1,000 of its generated web pages into image vectors and clustered them. The resulting design map has clear gaps — sitting exactly where purple gradients, bento-box layouts, and oversized headlines (the usual AI web clichés) would be.
  • Both models dodge these clichés, but for different reasons: GLM-5.2 simply never learned the tropes, while GPT-5.6 Sol learned them and refuses to render them.
  • It starts from a reliable template, then heavily customizes for each request — landing somewhere between cookie-cutter and total improvisation.
  • It didn't dodge everything: 26.5% of its outputs still stuff in confetti, and it's noticeably weak at building data charts.
1A counterintuitive finding

You can spot an "AI website" at a glance — this one learned not to look like one

You can usually spot an AI-made website in a second: a purple-to-blue gradient background, a screen chopped into bento-box cards, a giant headline hogging the whole hero section. See enough of them, and that unmistakably "fake" plastic sheen practically becomes an AI website's birthmark.

OpenAI's new model, GPT-5.6 Sol, recently took first place on third-party benchmark Design Arena's web design leaderboard — 18 spots above its predecessor, GPT-5.5, and the first time an OpenAI model has ever topped it. When the benchmark team dug into its outputs, they found its biggest edge was learning to actively avoid these instantly-recognizable clichés.

Design Arena's method: feed different models the same web-generation task, put the results in front of humans for blind selection, and rank models by how often they get picked. GPT-5.6 Sol won the "web design (non-agentic)" board — meaning single-shot generation, not multi-turn agentic revision.

Why it matters: Topping the chart is just the headline. What the benchmark team did next is more interesting — they converted its thousand outputs into image vectors and plotted them as a design map, offering a first direct look at the gaps left behind when a model knows what to draw and chooses not to.
#1
GPT-5.6 Sol's rank on the web design (non-agentic) board, 18 spots above its predecessor
1,000
GPT-5.6-generated web pages the benchmark team used to build the design map
GPT-5.6 Sol design benchmark overview
Design Arena's benchmark overview for GPT-5.6 Sol. Source: Design Arena
2First, what exactly is being avoided

The tells that scream "an AI made this"

The purple gradients and bento boxes mentioned above are just a couple of examples. In the field, this whole family of tics has a name: anti-patterns, or "design smells" — dead giveaways the moment you see them. To understand this finding, it helps to see the full lineup first.

Three months ago, when analyzing GPT-5.5, Design Arena put together a list of these design smells. The main ones:

Design smellWhat it looks like
Purple-to-blue gradientBackground uniformly washed in a purple-to-blue gradient — the signature AI-image color scheme
Bento-box layoutPage chopped into unevenly sized card blocks, packed into the screen like a bento box
Oversized headline textSkips the hero image, uses one giant line of headline text to fill the whole first screen
Offset layoutElements deliberately staggered left and right to look "intentionally designed" — really just a trope
Grid backgroundA faint grid line pattern laid over the entire page

None of these is wrong on its own — the problem is AI models overuse them, to the point readers can spot them instantly. In Design Arena's blind tests, outputs carrying these smells are usually the ones that lose.

losing design with purple blue gradient
Smell #1: purple-blue gradient. This kind of output is usually the loser in blind comparisons. Source: Design Arena
losing design with grid background
Smell #2: grid background. Source: Design Arena
3The hole in the design space

Plot a thousand web pages as a map, and a few patches go missing

Good taste and bad taste are both subjective calls. The benchmark team wanted evidence you could actually see, so they turned GPT-5.6 Sol's 1,000 generated pages into points on a map.

The process has two steps. First, every page screenshot is fed into a model called CLIP, which spits out a long string of numbers for each image — think of it as a "genetic code" for that image's style; images with similar styles get similar codes. Second, a technique called UMAP flattens that long code down into a single point on a 2D plane, while trying to preserve the original distances — pages with similar styles still land close together on the map.

Think of it this way

CLIP is like issuing each web page a "style ID card" — a long code. UMAP then flattens a three-dimensional nebula into a flat star chart; the clumps, the density, and the empty patches in the original cloud mostly survive the flattening.

Spread out, those thousand points form the model's "design space" — the whole region of style it likes to generate, generates often. Once plotted, the benchmark team found something that shouldn't be there: clear holes in the map, patches of the point cloud conspicuously missing.

GPT-5.5's design map Point cloud is continuous, no gaps Purple gradients, bento boxes and the rest all show up here, business as usual GPT-5.6 Sol's design map Same spread, but chunks are missing Purple gradient should be here, isn't Bento box
The two point clouds should look similar, but GPT-5.6 Sol (right) suddenly has no points in a few spots — the dashed circles mark the "holes." Illustrative diagram, drawn from the benchmark team's description

Below is the benchmark team's actual projection. The first image is GPT-5.6 Sol — dense with points, but with visible gaps; the second is GPT-5.5, whose point cloud is filled in, with no such gaps.

GPT-5.6 Sol UMAP projection with holes
GPT-5.6 Sol's UMAP projection, with visible holes in the point cloud. Source: Design Arena
GPT-5.5 UMAP projection filled
GPT-5.5 projected the same way — a filled-in point cloud, no such gaps. Source: Design Arena
What the holes mean

UMAP projections are designed to preserve holes from the original space. So when one model's map has a hole and another's doesn't, it means: that spot is somewhere GPT-5.6 Sol could generate — and does generate elsewhere — but chooses not to put anything there.

To confirm what's sitting inside the holes, the benchmark team overlaid GPT-5.6 Sol's and GPT-5.5's outputs on the same chart: GPT-5.6 Sol's points dyed orange, laid over GPT-5.5's. Wherever there are orange dots is territory GPT-5.6 Sol will visit; wherever there's only the base color and no orange, that's territory it avoids that the previous model would have gone to.

overlap of GPT-5.6 Sol orange and GPT-5.5 blue dots showing purple gradient cluster gap
Orange dots (GPT-5.6 Sol) overlaid on blue dots (GPT-5.5). The two mostly overlap, except for the "purple gradient" cluster, which has no orange at all — the one region GPT-5.6 Sol avoids completely. Source: Design Arena

The two models overlap almost everywhere, except in the purple-gradient cluster, where there isn't a single orange dot. Bento boxes, oversized headlines, offset layouts — same story. What's sitting in those holes is exactly that batch of AI clichés.

4Learned it, but refuses — versus never learned it at all

Two models dodge the same clichés — by two very different routes

GPT-5.6 Sol isn't the only model that avoids AI clichés. But how it avoids them is different from everyone else, and that's the most counterintuitive part of this whole analysis.

Take GLM-5.2, which ranks ahead of it, as a comparison. It also rarely produces oversized-headline-style clichés — but it does so by learning from a batch of high-scoring templates that simply never contained those tropes to begin with. In other words, its design space never had a "purple gradient" region to draw from, so naturally it can't render one; no hole shows up on its map, because that spot was empty from the start.

GPT-5.6 Sol is a different story. The fact that its map has holes is exactly what shows it learned these tropes and has the ability to render them — it just steers away every single time it generates. It knows what a purple gradient looks like and where it belongs, and then chooses not to put it there.

GPT-5.6 Sol — learned it, but refuses

Its design space covers where these clichés would sit, yet it actively steers away from generating there, leaving hole after hole on the map. This is "knows, and chooses not to."

GLM-5.2 — never learned it at all

It draws from a batch of templates that never contained these clichés — that style was never within its capability range to begin with, so the corresponding region never even exists on the map. This is "the option was never there."

GLM-5.2 avoids anti-patterns via templates
GLM-5.2 avoids clichés through its templates — its design space has no corresponding region, so no hole appears. Source: Design Arena

Both routes can end up avoiding the same clichés in the final product. The difference shows in the shape of the capability itself: one draws a circle around that territory and labels it "off-limits," the other never mapped that territory in the first place. The benchmark team argues GPT-5.6 Sol looks more like learning followed by active suppression — a rare trait among models.

5Somewhere between template and custom

Starts from a reliable template, then heavily reworks each request

This is the benchmark team's second finding, about "personalization." Models that build web pages roughly split into two camps: heavily template-reliant ones that stay consistent but tend to look the same every time, and ones that barely use templates at all, generating everything from scratch each time — more variety, less reliability. GPT-5.6 Sol lands in the middle.

It starts from a set of proven, good design structures, then makes substantial adjustments for each specific request, so a whole batch of variants can grow out of a shared template — related to each other, yet each distinct. The benchmark team used an analogy: like bacteria evolving into different closely related strains, sharing a common foundation before branching out on their own.

Heavily templated Consistent, but often repetitive Barely templated Custom every time, lots of variety GLM-5.2 Claude Fable 5 GPT-5.6 Sol Template base, heavily reworked
The template-reliance spectrum: GLM-5.2 sits toward the templated end, Claude Fable 5 toward the fully-custom end, and GPT-5.6 Sol lands in the middle, leaning custom. Drawn from the benchmark team's description

As reference points for the two ends: GLM-5.2 also ranks high this round, and it does so precisely by learning from a batch of high-scoring templates — the upside is consistency, the downside is that variation mostly comes only from the request itself, with the same template showing up over and over. Claude Fable 5, by contrast, shows almost no template fingerprint at all; its design space is more scattered, with every output heavily customized to the request.

GLM-5.2 templated designs look similar
GLM-5.2's outputs show a clear template fingerprint, and look similar to one another. Source: Design Arena
Claude Fable 5 varied non-templated designs
Claude Fable 5 shows almost no template fingerprint, and its design space is far more scattered and varied. Source: Design Arena

GPT-5.6 Sol sits between these two ends: templates keep the floor high, customization pulls the outputs apart. The benchmark team believes this lets every user get a page that fits their own needs while still looking like professional work — a major reason it ranks so high. One small tell: when picking images for different pages, it often reuses the same image across several different contexts.

6The clichés it didn't dodge

Confetti everywhere, and charts still need work

GPT-5.6 Sol doesn't nail active cliché-suppression across the board. The benchmark flags two clear weak spots.

First: confetti. It has a particular fondness for scattering confetti animations across pages — showing up in over 26.5% of its outputs. How bad is it: even when it isn't handed a ready-made confetti library, it'll just hand-write one and use it anyway. This is itself a well-known AI cliché — just one it didn't manage to dodge.

GPT-5.6 Sol overuses confetti
GPT-5.6 Sol overuses confetti, showing up in more than a quarter of its outputs. Source: Design Arena

Second: data charts. It's noticeably weaker at charts and data visualization, and can't produce a convincing real-world chart using chart.js (a common web charting library).

GPT-5.6 Sol weak at charts
GPT-5.6 Sol performs relatively poorly on charts and data visualization. Source: Design Arena
What it actively dodges

Purple gradients, bento boxes, oversized headlines, offset layouts, grid backgrounds — all leave holes on the map.

What it didn't dodge

Confetti shows up in 26.5% of outputs — and it'll hand-write its own confetti library; data charts remain unpolished.

7Faster and cheaper too

More than twice as fast as the previous leader, at half the price

Beyond rankings and taste, GPT-5.6 Sol also pulls ahead on speed and price. The benchmark team says it establishes two new efficient frontiers at once: one for "quality versus speed," one for "quality versus price."

2.44x faster
Generation speed versus the previous #1, GLM 5.2
36% faster
Generation speed versus Claude Fable 5
$5 / $30
Price per million tokens, input / output
$10 / $50
Claude Fable 5's price at the same rate — twice as expensive
What "Pareto frontier" means

Picture a drawn boundary line: standing on it, if you want a nicer-looking page, you have to spend more time or more money — there's no free lunch. GPT-5.6 Sol pushes that boundary forward, meaning it can hit the same level of visual quality faster and cheaper.

GPT-5.6 Sol pareto frontier preference vs speed and price
GPT-5.6 Sol establishes a new Pareto frontier on both the "quality vs. speed" and "quality vs. price" charts. Source: Design Arena
8What this means for picking a model

More discerning, and more flexible

The benchmark team boiled this down to two sentences.

First, it's more discerning. It appears to have learned which visual tropes make a web page instantly readable as "made by AI," and actively suppresses them, while keeping a set of reliable design structures ready to use. Second, it's more flexible. It fuses templates and customization together: templates set a reliable floor, then heavy per-request rework makes sure everyone gets a page that fits them, while still reading as professional work.

Taken together, these two points are Design Arena's main explanation for why GPT-5.6 Sol leads the single-shot leaderboard. As for the confetti it didn't dodge and the charts it can't quite pull off, those are laid out in the same benchmark — you need the whole picture, not just the highlight reel.

GLM-5.2 looks like it never learned these design smells in the first place, so it simply doesn't produce them; GPT-5.6 Sol looks like it learned them, knows these smells exist, but refuses to render them. Design Arena, analysis of GPT-5.6 Sol
Source: Design Arena, "We benchmarked GPT-5.6 Sol on web design," July 15, 2026. Rankings, image clustering analysis, and speed/price figures in this piece all come from Design Arena's own leaderboard and analysis; images from the same source. GPT-5.6 Sol, GPT-5.5, GLM-5.2, and Claude Fable 5 are model names belonging to their respective vendors.