Skip to content
GlanceHall
  • Builds
  • Models
  • Compare
  • How we test
Get the next round
  • Builds
  • Models
  • Compare
  • How we test
  • Journal
  • About
Get the next round

We don't rate. We let you see.

contact@glancehall.com

Watch

  • All builds
  • Models
  • Head to head
  • Journal

Trust

  • How we test
  • About
  • FAQ
  • Contact

Legal

  • Privacy
  • Terms
  • Cookies
© 2026 GlanceHallIndependent. Not affiliated with Anthropic, OpenAI or any model vendor.
Journal/Can AI build it?

Claude web design: what Opus 5.5 and Fable 5.1 designed from one brief

Is Claude good at web design? Claude Opus 5.5 and Claude Fable 5.1 designed a landing page, a dashboard, a phone app and an invented brand's site. What they share, and which Claude to use.

By GlanceHall · Published October 22, 2026 · 2 min read

Claude is the most distinctive web designer we have tested, and the two current Claude models design in recognisably different voices. Below are their pages for four briefs: a landing page, an analytics dashboard, a phone app and a website for a product each model invented. Same prompts as every other model, one attempt, no edits.

SaaS landing page · same prompt, one shot, unedited
Blind testOpen side by side
Claude Opus 5.5
Open Claude Opus 5.5’s build
Claude Fable 5.1
Open Claude Fable 5.1’s build

Claude Opus 5.5 (left) and Claude Fable 5.1 (right): the same landing page brief, two design voices.

How does Claude design a landing page?

As a product demo. Claude Opus 5.5 built a working booking calendar into the hero, with a floating "focus guard" card and buffer reminders, under "Easy to book. Hard to interrupt." in Fraunces and Geist. Claude Fable 5.1 went the other way: one large Instrument Serif headline, "Get booked without giving up your best hours", over a soft gradient, with an animated cursor booking and paying for a session below it. Both pages work on a phone and both pricing switches change every price.

How does Claude design an interface?

With annotation and warmth. On the dashboard brief both Claude models chose Inter with Instrument Serif and warm orange accents on white cards, and both added touches that help you read the data: Opus labels the revenue peak and draws the average line, Fable labels the best day and lets you flip the chart into a table.

Analytics dashboard · same prompt, one shot, unedited
Blind testOpen side by side
Claude Opus 5.5
Open Claude Opus 5.5’s build
Claude Fable 5.1
Open Claude Fable 5.1’s build

The two Claude dashboards: same data, same typefaces, different emphasis.

Detailed close-up of white text printed on black paper, showcasing typography.
Photo by Brett Jordan on Pexels.

How does Claude design a phone app?

Like a native app, with feedback for every step. On the phone finance brief, Opus 5.5 drew a sparkline of the month's balance inside the balance card and ended the send-money flow on a full-screen green confirmation reading "$42.50 sent". Fable 5.1 added money-in and money-out chips under the balance and confirmed the transfer with a "Sent!" overlay while the new balance counted down to $4,783.69. Both kept the brief's bottom tab bar and safe areas, and both updated the transaction list after sending.

Mobile finance app · same prompt, one shot, unedited
Blind testOpen side by side
Claude Opus 5.5
Open Claude Opus 5.5’s build
Claude Fable 5.1
Open Claude Fable 5.1’s build

Fern, a fictional finance app, as designed by the two Claude models. Recorded on a 390 by 844 phone viewport.

What does Claude design when nothing is specified?

Something with a story. Given a brief to invent a product and make its website memorable, Opus 5.5 created Nod, a sleeper train that loops through the night and returns you to the same platform, with a flip-clock headline and a section of synthesised train sound. Fable 5.1 created Yonder, a lamp that shows the sky over someone's faraway city, on a dusky page with a glowing globe. Both chose night, distance and a single emotional idea, and both set their pages in Fraunces.

Open brief · same prompt, one shot, unedited
Blind testOpen side by side
Claude Opus 5.5
Open Claude Opus 5.5’s build
Claude Fable 5.1
Open Claude Fable 5.1’s build

Claude's two inventions for the open brief.

Which Claude should you use for web design?

Opus 5.5 when the page has to sell a product by showing it; Fable 5.1 when it has to set a mood. Opus is cheaper per token, $4 and $20 per million against Fable's $10 and $50 on Anthropic's pricing page, though it works longer and so costs about the same per page. Both took between 25 and 74 minutes per site on our briefs, against 9 to 13 minutes for the GPT models. To compare Claude's designs with GPT's on the same briefs, see best AI for web design.

Can you spot the Claude designs without labels? Four sites, names hidden.

Take the blind test
Close-up image showcasing various styles of the letter A in print, emphasizing typography diversity.
Photo by Brett Jordan on Pexels.

Frequently asked questions

Is Claude good for website design?

Yes. Claude Opus 5.5 and Claude Fable 5.1 produced the most distinctive designs of the four models we tested, with working interactions and responsive layouts, from one prompt and with no edits. They took longer than the GPT models: 25 to 74 minutes per site.

What Claude model is best for web design?

Claude Opus 5.5 for rich, product-led pages with interactive demos; Claude Fable 5.1 for calmer, editorial pages with fewer elements and more space. Opus is also the cheaper of the two per token.

Can Claude design a website without code?

Claude writes the code for you. In the Claude app you can preview a page as an Artifact on any plan; in Claude Code, available from the Pro plan, it writes the files, opens the page in a browser and refines it, which is how these designs were made.

Do Claude's two models design differently?

On the landing page, clearly: Opus put a working booking calendar in a dense hero, Fable a single large serif headline over a quiet gradient. On the dashboard they converged, with the same typefaces and similar warm orange accents.

Judge the builds yourself.

Eight briefs, four models, nothing edited. Watch them side by side or take the blind test.

Watch the buildsHow we test

Keep reading

Can AI build it?Claude as a website builder: the site it built from one prompt (open it)Can Claude build a website? Claude Opus 5.5 and Claude Fable 5.1 each built the same marketing site from one prompt. Open both, see what worked, and learn which Claude to use.October 22, 2026Best AI for…AI web design in 2026: four models, one brief, and what good looks like nowWhat AI web design produces in 2026, shown through four landing pages built by frontier models from one brief: the conventions they share, where they succeed, and how to get more.October 22, 2026Best AI for…Best AI for web design: four landing pages from one briefWhich AI is best for web design? Instead of a tool list, here are four landing pages built from the same prompt by Claude Opus 5.5, Claude Fable 5.1, GPT-6 Astra and GPT-6.1 Sol, unedited.October 22, 2026