Arena WebDev: Results as Published

Code Arena, earlier called WebDev Arena, is a live human-preference leaderboard for building web applications. A user submits a prompt, such as building a simple chess game, two anonymous models each generate a web app, the user interacts with both and votes for the better one (ties allowed). The votes feed a Bradley-Terry model that estimates each model's strength from pairwise results, printed as an Elo-style rating. The scope has grown from single pages to multi-file applications, dashboards, browser games and editing tools.

Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.