evals.report
BenchmarksLabsCompareRun guides
BenchmarksChat preference

Design Arena

A crowdsourced human-preference benchmark where top AI models receive identical design/frontend prompts and users vote head-to-head on the anonymized outputs, producing a Bradley-Terry (Elo) ranking of design taste across categories like websites, UI components, games, and data visualization.

Chat preferenceEloHigher is better
ScoresAboutRun this benchmark
ModelLabScore↓Source modelStatusDate
Claude Opus 4.6Anthropic1344—VerifiedFeb 5, 2026Details
GLM-5.1Z.ai1335—VerifiedApr 7, 2026Details
Kimi K2.6Moonshot AI1335—VerifiedApr 20, 2026Details
Claude Opus 4.7Anthropic1328—VerifiedApr 16, 2026Details
Claude Sonnet 4.6Anthropic1327—VerifiedFeb 17, 2026Details
MiMo-V2.5-ProXiaomi1325—VerifiedApr 22, 2026Details
MiniMax M3MiniMax1321—VerifiedJun 1, 2026Details
MiMo-V2.5Xiaomi1309—VerifiedApr 22, 2026Details
Muse SparkMeta1306—VerifiedApr 8, 2026Details
DeepSeek V4 ProDeepSeek1302—VerifiedApr 24, 2026Details
GPT-5.5OpenAI1301—VerifiedApr 23, 2026Details
GLM-5Z.ai1300—VerifiedFeb 11, 2026Details
Gemini 3.5 FlashGoogle DeepMind1297—VerifiedMay 19, 2026Details
Claude Opus 4.5Anthropic1295—VerifiedNov 24, 2025Details
Gemini 3 ProGoogle DeepMind1295—VerifiedNov 18, 2025Details
Kimi K2.5Moonshot AI1292—VerifiedJan 27, 2026Details
MiniMax M2.7MiniMax1285—VerifiedMar 18, 2026Details
Claude Opus 4.8Anthropic1282—VerifiedMay 28, 2026Details
Gemini 3.1 Pro PreviewGoogle DeepMind1282—VerifiedFeb 19, 2026Details
Qwen 3.6 PlusAlibaba / Qwen1281—VerifiedApr 2, 2026Details
GLM-4.7Z.ai1273—VerifiedDec 22, 2025Details
Grok 4.20 beta reasoningxAI1271—VerifiedMar 9, 2026Details
DeepSeek V4 FlashDeepSeek1268—VerifiedApr 24, 2026Details
GPT-5.4OpenAI1264—VerifiedMar 5, 2026Details
MiniMax M2.5MiniMax1261—VerifiedFeb 12, 2026Details
Grok 4.3xAI1260—VerifiedApr 17, 2026Details
MiniMax M2.1MiniMax1245—VerifiedDec 23, 2025Details
Gemini 3 FlashGoogle DeepMind1244—VerifiedDec 17, 2025Details
Claude Sonnet 4.5Anthropic1235—VerifiedSep 29, 2025Details
Qwen3.5-397B-A17BAlibaba / Qwen1233—VerifiedFeb 16, 2026Details
Claude 3.7 SonnetAnthropic1231—VerifiedFeb 24, 2025Details
GPT-5.2OpenAI1224—VerifiedDec 11, 2025Details
GPT-5OpenAI1223—VerifiedAug 7, 2025Details
DeepSeek V3.2DeepSeek1220—VerifiedDec 1, 2025Details
GLM-4.6Z.ai1220—VerifiedSep 30, 2025Details
Claude Opus 4.1Anthropic1219—VerifiedAug 5, 2025Details
GPT-5.1OpenAI1216—VerifiedNov 12, 2025Details
Claude Opus 4Anthropic1215—VerifiedMay 22, 2025Details
Gemini 2.5 ProGoogle DeepMind1208—VerifiedMar 25, 2025Details
GPT-5.3-CodexOpenAI1199—VerifiedFeb 5, 2026Details
Qwen 3 Coder 480BAlibaba / Qwen1197—VerifiedJul 22, 2025Details
Claude Sonnet 4Anthropic1196—VerifiedMay 22, 2025Details
DeepSeek R1DeepSeek1193—VerifiedJan 20, 2025Details
Mistral Medium 3.5Mistral AI1176—VerifiedApr 28, 2026Details
GPT-5 miniOpenAI1170—VerifiedAug 7, 2025Details
Claude Haiku 4.5Anthropic1169—VerifiedOct 15, 2025Details
DeepSeek V3.1DeepSeek1166—VerifiedAug 21, 2025Details
Qwen3 MaxAlibaba / Qwen1165—VerifiedSep 5, 2025Details
DeepSeek V3 0324DeepSeek1163—VerifiedMar 24, 2025Details
Grok 4.1 fast reasoningxAI1142—VerifiedNov 19, 2025Details
Gemini 2.5 FlashGoogle DeepMind1113—VerifiedApr 17, 2025Details
Qwen3 235B A22B Instruct 2507Alibaba / Qwen1093—VerifiedJul 21, 2025Details
Kimi K2 InstructMoonshot AI1088—VerifiedJul 11, 2025Details
GPT-4.1OpenAI1080—VerifiedApr 14, 2025Details
o3OpenAI1074—VerifiedApr 16, 2025Details
Grok 4xAI1070—VerifiedJul 9, 2025Details
o4-miniOpenAI1030—VerifiedApr 16, 2025Details
OLMo 3.1-Think 32BAllen Institute for AI1029—VerifiedDec 12, 2025Details
GPT-OSS-120BOpenAI1017—VerifiedAug 5, 2025Details
Llama 4 MaverickMeta934—VerifiedApr 5, 2025Details
GPT-4oOpenAI915—VerifiedMay 13, 2024Details
Llama 4 ScoutMeta844—VerifiedApr 5, 2025Details

Each row reports the model’s Elo on Design Arena. Click a row for the full run context.