Atlas

Benchmarks

← All benchmarks

Arena Code WebDev — Simulations

Code

Arena Code WebDev rating for interactive simulations. Anonymous generated simulations are compared through blind pairwise user votes.

Top models (higher is better)

ModelScore
Claude Opus 5.51922
GPT-6 Astra1883
Claude Sonnet 5.51841
GPT-6.1 Sol1829
Claude Fable 5.11818
Muse Wonder Full1782
Claude Opus 51751
Kimi K3.1 Pro1745
Gemini 4 Argon1742
GPT-6 Sol1741
Hy4 Preview1703
DeepSeek V4.1 Flash 10171701
Qwen3.8 Max (0902)1698
MiMo-V2.6-Pro1691
Grok 4.71690
Muse Spark 1.31690
Muse Wonder1688
Kimi K31684
Qwen3.8-Max1676
Claude Fable 51673
MiMo-V2.6-Flash1660
GLM-5.31656
Grok 4.61648
Qwen3.8-Flash-Next1644
GLM-5.21635
DeepSeek V4.1 Flash1633
Claude Haiku 5.51632
GPT-5.6 Sol1626
GLM-5.3 Flash1620
DeepSeek V4 Pro 08131619
Gemini 3.7 Flash1618
GPT-6 Luna1610
Gemini 3.8 Flash1607
Qwen3.8 27B1598
MiniMax M3.1 Flash1590
Opus 4.81576
Grok 4.51575
Opus 4.71573
DeepSeek-V4-Flash1568
Opus 4.61563
Loading Atlas data…