- Cleus
- AI benchmarks
- WebDev Arena
WebDev Arenaexplained
People vote on which AI built the better website. Someone types a request, for example “build a to-do app with dark mode”, and two anonymous models each build a working web page. The person tries both and votes for the better one. Thousands of these votes turn into a rating, like chess Elo.
- Highest published score
- Claude Opus 5.5, 1,820
- Leader among popular models
- Claude Opus 5.5
- Models with a result
- 105
- Skill it tests
- Coding
What WebDev Arena really tests
- What it is
- Someone types a request, for example “build a to-do app with dark mode”, and two anonymous models each build a working web page. The person tries both and votes for the better one. Thousands of these votes turn into a rating, like chess Elo.
- A question like
- “Make a landing page for a coffee shop with a menu, opening hours and a booking form.” Two finished pages appear side by side; you pick one. Our own example in the style of the test, not a real item.
- Why it matters
- It is the most direct measure of “which AI makes the nicest working website or app from one prompt”, judged by real people instead of a script.
- Reading the score
- An Elo-style rating. Higher is better; 100 points apart means the higher model wins roughly 64% of head-to-head votes.
- The catch
- Voters reward looks and polish, not only whether the code is correct. Ratings are relative: a gap of under about 20 points is within the margin of error.
- Made by
- LMArena. Official page
Leaderboard
Each model’s best published result across its reasoning settings.
Popular models
The models we track across this site.
Top 20 of all 105 models tested
Includes older, preview and specialist models that aren’t on our main leaderboard.
Questions
What is WebDev Arena?
Someone types a request, for example “build a to-do app with dark mode”, and two anonymous models each build a working web page. The person tries both and votes for the better one. Thousands of these votes turn into a rating, like chess Elo.
What does a WebDev Arena score mean?
An Elo-style rating. Higher is better; 100 points apart means the higher model wins roughly 64% of head-to-head votes.
Which AI model is best on WebDev Arena?
Claude Opus 5.5 from Anthropic has the highest published WebDev Arena result we have (1,820).
Is WebDev Arena a good benchmark?
It is the most direct measure of “which AI makes the nicest working website or app from one prompt”, judged by real people instead of a script. The catch: Voters reward looks and polish, not only whether the code is correct. Ratings are relative: a gap of under about 20 points is within the margin of error.
Who made WebDev Arena?
LMArena.
Benchmark scores: Epoch AI, Capabilities & benchmarking, used under CC BY 4.0. For each model we show its best published result across reasoning settings. Prices and context sizes: OpenRouter. Data refreshed November 9, 2026. Hands-on tests (answers, websites, code, refusals, speed) are our own single runs. Explanations written by Cleus.