

Which model can create the best website from a prompt?


Claude Fable 5.2
Claude Opus 5.5

GPT-6 Astra
(xhigh)

Kimi K3.1 Pro
Claude Fable 5.1

Kimi K3

Muse Spark 1.3 Max
Claude Opus 5

Qwen3.8 Max
Claude Fable 5
Grok 4.6

GPT-5.6 Sol
(xhigh)

GPT-5.3 Codex
Evaluates model ability to build complete React applications with authentication, databases, and backend functionality from real-world user prompts. In addition to frontend agent behavior, the evaluation captures use of backend tools such as supabase_create_table, supabase_insert, supabase_query, supabase_update, supabase_delete, supabase_auth_create_user, supabase_create_bucket, supabase_create_rls_policy.


















Updated Nov 16, 2026, 1:20 AM UTC







Elo Rating: A skill-based rating system starting at 1200. Ratings are calculated using the standard Bradley-Terry model. See more in our Methodology page.Margin of Error: Shows ±standard error from the Bradley-Terry model for Elo rating confidence intervals.