llm.ing

Design Arena — pdf

mirrorrun by Design Arenaunit: elo34 modelssource ↗

Caveat: Head-to-head human votes on generated designs/code, mirrored via OpenRouter.

websitecodecategoriesgamedev3duicomponentdatavizsvgasciiartmobileappswebapps+19 more
#ModelScoreObserved
1Claude Fable 5.2
1348
Nov 19, 2026
2Claude Opus 5.5
1315
Nov 19, 2026
3Gemini 4 Carbon
1305
Nov 19, 2026
4Kimi K3
1276
Nov 19, 2026
5Gemini 4 Argon
1269
Nov 19, 2026
6MiniMax-M3
1264
Nov 18, 2026
7GLM-5.2
1255
Nov 19, 2026
8Claude Opus 5
1250
Nov 19, 2026
9Claude Sonnet 5.5
1248
Nov 19, 2026
10Grok 4.6
1238
Nov 19, 2026
11Kimi K3.1 Pro
1226
Nov 19, 2026
12Gemini 4 Hydrogen
1224
Nov 19, 2026
13Claude Sonnet 5
1217
Nov 19, 2026
14Claude Haiku 5.5
1215
Nov 19, 2026
15Gemini 3.8 Flash
1214
Nov 19, 2026
16Muse Spark 1.3
1206
Nov 19, 2026
17GLM-5.1
1194
Nov 18, 2026
18Grok 4.7
1193
Nov 18, 2026
19Claude Opus 4.6
1187
Nov 19, 2026
20Claude Opus 4.8
1181
Nov 19, 2026
21Gemini 3.6 Flash
1169
Nov 19, 2026
22Qwen3.7 Max
1163
Nov 19, 2026
23Muse Spark 1.1
1157
Nov 19, 2026
23Qwen3.6 Plus
1157
Nov 19, 2026
25Step 3.7 Flash
1152
Nov 19, 2026
26Gemini 3.7 Flash
1151
Nov 19, 2026
27Nemotron 3 Ultra 550B A55B
1149
Nov 19, 2026
28Gemini 3.5 Flash
1144
Nov 19, 2026
29Grok 4.5
1132
Nov 19, 2026
30Muse Spark 1.2
1130
Nov 19, 2026
31GPT-6.1 Sol
1123
Nov 19, 2026
32Qwen3.7 Plus
1120
Nov 19, 2026
33Gemini 3.1 Pro Preview
1098
Nov 19, 2026
34Grok 4.3
945
Nov 19, 2026

Dot = elo on a 924–1369 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.