llm.ing

Design Arena — webapps

mirrorrun by Design Arenaunit: elo33 modelssource ↗

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

websitecodecategoriesdatavizuicomponentgamedev3dsvgasciiartwebappsfullstack+15 more
#ModelScoreObserved
1Claude Fable 5
1317
Jul 24, 2026
2Kimi K3
1313
Jul 24, 2026
3Claude Opus 4.7
1311
Jul 23, 2026
4Claude Sonnet 5
1303
Jul 23, 2026
5Claude Opus 4.8
1281
Jul 23, 2026
6Grok 4.5
1275
Jul 24, 2026
7GLM-5.2
1274
Jul 23, 2026
8Kimi K2.6
1268
Jul 23, 2026
9Muse Spark 1.1
1263
Jul 23, 2026
10Claude Opus 4.6
1262
Jul 23, 2026
10MiniMax-M3
1262
Jul 23, 2026
12Gemini 3.5 Flash
1254
Jul 23, 2026
13Qwen3.7 Max
1253
Jul 23, 2026
14Claude Sonnet 4.6
1245
Jul 23, 2026
15Kimi K2.7 Code
1233
Jul 23, 2026
16GLM-5.1
1229
Jul 23, 2026
17Claude Opus 4.5 (latest)
1216
Jul 23, 2026
18Grok 4.20 (Non-Reasoning)
1201
Jul 24, 2026
19Gemini 3.1 Pro Preview
1183
Jul 23, 2026
20Kimi K2.5
1181
Jul 23, 2026
21Grok 4.3
1180
Jul 24, 2026
22GPT-5.5
1171
Jul 24, 2026
22Gemini 3 Flash Preview
1171
Jul 23, 2026
24GLM-5V-Turbo
1169
Jul 24, 2026
25GPT-5.2
1142
Jul 23, 2026
26GPT-5-Codex
1138
Jul 23, 2026
27Claude Sonnet 4.5 (latest)
1129
Jul 23, 2026
28GPT-5.4
1122
Jul 23, 2026
29GPT-5.2 Codex
1111
Jul 23, 2026
30GPT-5.3 Codex
1103
Jul 23, 2026
31GPT-5.1 Codex
1076
Jul 23, 2026
32GPT-5.1
1075
Jul 23, 2026
33DeepSeek V4 Pro
1005
Jul 23, 2026

Dot = elo on a 989–1333 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.