Design Arena Elo
Design Arena gives the same prompt to several models and shows voters two of the outputs side by side, without model names. Each choice is one pairwise vote. The votes are turned into a rating with a Bradley-Terry model, shown as an Elo-style number that starts at 1200, with a separate leaderboard per category such as website, UI component, SVG, data visualization, 3D and game development. A rating reflects human preference on visual and interactive output. It depends on who votes and which prompts are used, and ratings from different categories are not on the same scale.
Sources: Design Arena: how the rankings work
Open-weight models
| Model | Category | Endpoint | Win rate | |
|---|---|---|---|---|
| Kimi K3 | 3d | moonshotai/kimi-k3-20260715 | 1394 | 67.7% |
| GLM 5.3 | 3d | z-ai/glm-5.3-20260816 | 1345 | 60.7% |
| GLM 5.3 Flash | 3d | z-ai/glm-5.3-flash-20260826 | 1321 | 56.5% |
| DeepSeek V4 Pro | 3d | deepseek/deepseek-v4-pro-20260423 | 1257 | 56.8% |
| Kimi K2.7 Code | 3d | moonshotai/kimi-k2.7-code-20260612 | 1245 | 50.5% |
| MiniMax M3 | 3d | minimax/minimax-m3-20260531 | 1212 | 50.2% |
| Nemotron 3 Ultra | 3d | nvidia/nemotron-3-ultra-550b-a55b-20260604 | 1135 | 39.2% |
| Step 3.7 Flash | 3d | stepfun/step-3.7-flash-20260528 | 1132 | 40.6% |
| GLM 5.3 Flash | asciiart | z-ai/glm-5.3-flash-20260826 | 1272 | 55.0% |
| GLM 5.3 | asciiart | z-ai/glm-5.3-20260816 | 1207 | 46.1% |
| Kimi K2.7 Code | asciiart | moonshotai/kimi-k2.7-code-20260612 | 1205 | 48.6% |
| MiniMax M3 | asciiart | minimax/minimax-m3-20260531 | 1169 | 46.8% |
| Step 3.7 Flash | asciiart | stepfun/step-3.7-flash-20260528 | 1158 | 44.9% |
| DeepSeek V4 Pro | asciiart | deepseek/deepseek-v4-pro-20260423 | 1155 | 46.4% |
| Nemotron 3 Ultra | asciiart | nvidia/nemotron-3-ultra-550b-a55b-20260604 | 1104 | 38.1% |
| Kimi K3 | codecategories | moonshotai/kimi-k3-20260715 | 1367 | 62.8% |
| GLM 5.3 | codecategories | z-ai/glm-5.3-20260816 | 1312 | 54.7% |
| GLM 5.3 Flash | codecategories | z-ai/glm-5.3-flash-20260826 | 1287 | 49.4% |
| Kimi K2.7 Code | codecategories | moonshotai/kimi-k2.7-code-20260612 | 1257 | 50.0% |
| MiniMax M3 | codecategories | minimax/minimax-m3-20260531 | 1252 | 50.9% |
| DeepSeek V4 Pro | codecategories | deepseek/deepseek-v4-pro-20260423 | 1244 | 51.9% |
| Step 3.7 Flash | codecategories | stepfun/step-3.7-flash-20260528 | 1186 | 42.6% |
| Nemotron 3 Ultra | codecategories | nvidia/nemotron-3-ultra-550b-a55b-20260604 | 1149 | 36.1% |
| Kimi K3 | dataviz | moonshotai/kimi-k3-20260715 | 1354 | 63.2% |
| GLM 5.3 | dataviz | z-ai/glm-5.3-20260816 | 1269 | 52.7% |
| GLM 5.3 Flash | dataviz | z-ai/glm-5.3-flash-20260826 | 1266 | 50.3% |
| MiniMax M3 | dataviz | minimax/minimax-m3-20260531 | 1239 | 51.3% |
| Kimi K2.7 Code | dataviz | moonshotai/kimi-k2.7-code-20260612 | 1231 | 48.8% |
| DeepSeek V4 Pro | dataviz | deepseek/deepseek-v4-pro-20260423 | 1206 | 48.7% |
| Step 3.7 Flash | dataviz | stepfun/step-3.7-flash-20260528 | 1176 | 43.0% |
| Nemotron 3 Ultra | dataviz | nvidia/nemotron-3-ultra-550b-a55b-20260604 | 1144 | 37.4% |
| Kimi K3 | gamedev | moonshotai/kimi-k3-20260715 | 1369 | 61.4% |
| GLM 5.3 | gamedev | z-ai/glm-5.3-20260816 | 1320 | 52.8% |
| GLM 5.3 Flash | gamedev | z-ai/glm-5.3-flash-20260826 | 1288 | 47.2% |
| DeepSeek V4 Pro | gamedev | deepseek/deepseek-v4-pro-20260423 | 1236 | 52.3% |
| Kimi K2.7 Code | gamedev | moonshotai/kimi-k2.7-code-20260612 | 1225 | 47.0% |
| MiniMax M3 | gamedev | minimax/minimax-m3-20260531 | 1218 | 46.0% |
| Step 3.7 Flash | gamedev | stepfun/step-3.7-flash-20260528 | 1165 | 38.9% |
| Nemotron 3 Ultra | gamedev | nvidia/nemotron-3-ultra-550b-a55b-20260604 | 1144 | 36.4% |
| Kimi K3 | svg | moonshotai/kimi-k3-20260715 | 1305 | 59.8% |
| GLM 5.3 | svg | z-ai/glm-5.3-20260816 | 1294 | 56.8% |
| GLM 5.3 Flash | svg | z-ai/glm-5.3-flash-20260826 | 1288 | 57.1% |
| Kimi K2.7 Code | svg | moonshotai/kimi-k2.7-code-20260612 | 1175 | 44.4% |
| MiniMax M3 | svg | minimax/minimax-m3-20260531 | 1162 | 45.2% |
| DeepSeek V4 Pro | svg | deepseek/deepseek-v4-pro-20260423 | 1139 | 45.4% |
| Step 3.7 Flash | svg | stepfun/step-3.7-flash-20260528 | 1081 | 37.2% |
| Nemotron 3 Ultra | svg | nvidia/nemotron-3-ultra-550b-a55b-20260604 | 1068 | 32.5% |
| Kimi K3 | uicomponent | moonshotai/kimi-k3-20260715 | 1356 | 61.4% |
| GLM 5.3 Flash | uicomponent | z-ai/glm-5.3-flash-20260826 | 1324 | 55.0% |
| GLM 5.3 | uicomponent | z-ai/glm-5.3-20260816 | 1321 | 57.4% |
| Kimi K2.7 Code | uicomponent | moonshotai/kimi-k2.7-code-20260612 | 1265 | 51.0% |
| MiniMax M3 | uicomponent | minimax/minimax-m3-20260531 | 1250 | 50.9% |
| DeepSeek V4 Pro | uicomponent | deepseek/deepseek-v4-pro-20260423 | 1225 | 50.6% |
| Step 3.7 Flash | uicomponent | stepfun/step-3.7-flash-20260528 | 1185 | 42.2% |
| Nemotron 3 Ultra | uicomponent | nvidia/nemotron-3-ultra-550b-a55b-20260604 | 1145 | 36.4% |
| Kimi K3 | website | moonshotai/kimi-k3-20260715 | 1339 | 59.5% |
| GLM 5.3 | website | z-ai/glm-5.3-20260816 | 1303 | 54.6% |
| GLM 5.3 Flash | website | z-ai/glm-5.3-flash-20260826 | 1276 | 48.5% |
| Kimi K2.7 Code | website | moonshotai/kimi-k2.7-code-20260612 | 1268 | 51.4% |
| MiniMax M3 | website | minimax/minimax-m3-20260531 | 1262 | 51.9% |
| DeepSeek V4 Pro | website | deepseek/deepseek-v4-pro-20260423 | 1239 | 50.5% |
| Step 3.7 Flash | website | stepfun/step-3.7-flash-20260528 | 1199 | 44.2% |
| Nemotron 3 Ultra | website | nvidia/nemotron-3-ultra-550b-a55b-20260604 | 1145 | 34.9% |
Measured by Design Arena on hosted endpoints, not by this site. A quantized file you run yourself may score differently. One model can have several endpoints. Source: Design Arena (www.designarena.ai) via OpenRouter (openrouter.ai/rankings). designarena.ai