Atlas

Benchmarks

← All benchmarks

Arena Code WebDev — Brand & Marketing

Code

Arena Code WebDev rating for brand, marketing, and informational websites. Anonymous generated experiences are compared through blind pairwise user votes.

Top models (higher is better)

ModelScore
Claude Opus 5.51804
Claude Sonnet 5.51761
Claude Fable 5.11737
GPT-6 Astra1732
Muse Wonder Full1721
GPT-6.1 Sol1707
Kimi K3.1 Pro1704
Claude Opus 51681
Kimi K31678
Qwen3.8-Max1670
Qwen3.8 Max (0902)1667
DeepSeek V4.1 Flash 10171667
GPT-6 Sol1657
Muse Wonder1653
Grok 4.71645
GLM-5.31644
Muse Spark 1.31644
Qwen3.8-Flash-Next1634
MiMo-V2.6-Flash1632
MiMo-V2.6-Pro1627
GLM-5.3 Flash1623
Claude Fable 51623
GPT-5.6 Sol1622
Grok 4.61621
Gemini 4 Argon1620
GLM-5.21619
DeepSeek V4.1 Flash1618
Hy4 Preview1615
Qwen3.8 27B1608
DeepSeek-V4-Flash1607
MiniMax M3.1 Flash1594
DeepSeek V4 Pro 08131593
GPT-6 Luna1575
Mistral Large 4 Preview1570
Opus 4.71565
Gemini 3.8 Flash1563
Claude Haiku 5.51557
Sonnet 51557
Grok 4.51554
Gemini 3.7 Flash1553
Loading Atlas data…