Design Arena — codecategories
Caveat: Head-to-head human votes on generated designs/code, mirrored via OpenRouter.
| # | Model | Score | Observed | |
|---|---|---|---|---|
| 1 | Kimi K3Moonshot AI | 1431 | Jul 24, 2026 | |
| 2 | GLM-5.2Zhipu AI | 1341 | Jul 23, 2026 | |
| 3 | Claude Fable 5Anthropic | 1339 | Jul 23, 2026 | |
| 4 | Claude Opus 4.6Anthropic | 1326 | Jul 23, 2026 | |
| 5 | Gemini 3.6 FlashGoogle | 1322 | Jul 23, 2026 | |
| 6 | Claude Opus 4.7Anthropic | 1320 | Jul 24, 2026 | |
| 7 | Grok 4.5xAI | 1314 | Jul 24, 2026 | |
| 8 | Claude Sonnet 5Anthropic | 1312 | Jul 23, 2026 | |
| 9 | Kimi K2.6Moonshot AI | 1310 | Jul 24, 2026 | |
| 9 | GLM-5.1Zhipu AI | 1310 | Jul 24, 2026 | |
| 11 | Claude Sonnet 4.6Anthropic | 1308 | Jul 24, 2026 | |
| 12 | Muse Spark 1.1Meta | 1303 | Jul 24, 2026 | |
| 13 | Qwen3.7 MaxAlibaba | 1299 | Jul 23, 2026 | |
| 14 | Kimi K2.7 CodeMoonshot AI | 1291 | Jul 23, 2026 | |
| 15 | Qwen3.7 PlusAlibaba | 1290 | Jul 23, 2026 | |
| 16 | Gemini 3.5 FlashGoogle | 1289 | Jul 23, 2026 | |
| 17 | MiniMax-M3MiniMax (minimax.io) | 1287 | Jul 23, 2026 | |
| 18 | GPT-5.5OpenAI | 1284 | Jul 24, 2026 | |
| 19 | GLM-5Zhipu AI | 1279 | Jul 23, 2026 | |
| 20 | Claude Opus 4.5 (latest)Anthropic | 1275 | Jul 24, 2026 | |
| 21 | Gemini 3.1 Pro PreviewGoogle | 1274 | Jul 23, 2026 | |
| 22 | DeepSeek V4 ProDeepSeek | 1270 | Jul 24, 2026 | |
| 22 | Claude Opus 4.8Anthropic | 1270 | Jul 23, 2026 | |
| 22 | Kimi K2.5Moonshot AI | 1270 | Jul 23, 2026 | |
| 25 | MiniMax-M2.7MiniMax (minimax.io) | 1266 | Jul 23, 2026 | |
| 26 | GLM-5V-TurboZhipu AI | 1261 | Jul 23, 2026 | |
| 27 | Qwen3.6 PlusAlibaba | 1256 | Jul 24, 2026 | |
| 28 | Grok 4.20 (Non-Reasoning)xAI | 1251 | Jul 23, 2026 | |
| 29 | GLM-4.7Zhipu AI | 1249 | Jul 23, 2026 | |
| 30 | GPT-5.4OpenAI | 1244 | Jul 24, 2026 | |
| 31 | MiniMax-M2.5MiniMax (minimax.io) | 1240 | Jul 23, 2026 | |
| 32 | DeepSeek V4 FlashDeepSeek | 1238 | Jul 24, 2026 | |
| 33 | MiniMax-M2.1MiniMax (minimax.io) | 1223 | Jul 24, 2026 | |
| 34 | Gemini 3 Flash PreviewGoogle | 1222 | Jul 24, 2026 | |
| 35 | Grok 4.3xAI | 1221 | Jul 23, 2026 | |
| 36 | Claude Sonnet 4.5 (latest)Anthropic | 1215 | Jul 24, 2026 | |
| 37 | Qwen3.5 397B-A17BAlibaba | 1212 | Jul 24, 2026 | |
| 38 | GLM-4.7-FlashZhipu AI | 1210 | Jul 24, 2026 | |
| 39 | Claude Opus 4.1 (latest)Anthropic | 1204 | Jul 23, 2026 | |
| 39 | GPT-5.1OpenAI | 1204 | Jul 23, 2026 | |
| 41 | GPT-5.2OpenAI | 1202 | Jul 24, 2026 | |
| 42 | GPT-5OpenAI | 1201 | Jul 24, 2026 | |
| 43 | Qwen3.5 PlusAlibaba | 1200 | Jul 23, 2026 | |
| 43 | Step 3.7 FlashStepFun (China) | 1200 | Jul 24, 2026 | |
| 45 | GLM-4.6Zhipu AI | 1199 | Jul 23, 2026 | |
| 46 | GLM-4.5Zhipu AI | 1198 | Jul 23, 2026 | |
| 47 | DeepSeek V3.2DeepSeek | 1196 | Jul 24, 2026 | |
| 48 | Gemini 2.5 ProGoogle | 1187 | Jul 23, 2026 | |
| 49 | GPT-5.1 CodexOpenAI | 1183 | Jul 24, 2026 | |
| 50 | GPT-5.3 CodexOpenAI | 1177 | Jul 24, 2026 | |
| 51 | DeepSeek R1DeepSeek | 1171 | Jul 24, 2026 | |
| 52 | GLM-4.5-AirZhipu AI | 1170 | Jul 24, 2026 | |
| 53 | MiniMax-M2MiniMax (minimax.io) | 1168 | Jul 23, 2026 | |
| 54 | Nemotron 3 Ultra 550B A55BNvidia | 1157 | Jul 23, 2026 | |
| 55 | Claude Haiku 4.5 (latest)Anthropic | 1147 | Jul 24, 2026 | |
| 55 | GPT-5 MiniOpenAI | 1147 | Jul 24, 2026 | |
| 57 | Qwen3 MaxAlibaba | 1142 | Jul 24, 2026 | |
| 58 | DeepSeek ChatDeepSeek | 1141 | Jul 24, 2026 | |
| 59 | Gemini 2.5 FlashGoogle | 1137 | Jul 23, 2026 | |
| 60 | Kimi K2 ThinkingMoonshot AI | 1131 | Jul 23, 2026 | |
| 61 | GPT-5.1 Codex miniOpenAI | 1126 | Jul 24, 2026 | |
| 62 | GPT-5 NanoOpenAI | 1117 | Jul 24, 2026 | |
| 63 | Gemini 3.1 Flash Lite PreviewGoogle | 1104 | Jul 24, 2026 | |
| 64 | Qwen3 235B-A22BAlibaba | 1071 | Jul 24, 2026 | |
| 65 | GPT-4.1OpenAI | 1058 | Jul 24, 2026 | |
| 66 | o3OpenAI | 1053 | Jul 23, 2026 | |
| 67 | GPT-4.1 miniOpenAI | 1026 | Jul 24, 2026 | |
| 67 | Mercury 2Inception | 1026 | Jul 23, 2026 | |
| 69 | o4-miniOpenAI | 1009 | Jul 23, 2026 | |
| 70 | GPT-4.1 nanoOpenAI | 996 | Jul 23, 2026 | |
| 71 | GPT-4oOpenAI | 893 | Jul 24, 2026 |
Dot = elo on a 866–1458 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.