llm.ing

Design Arena — agenticslides

mirrorrun by Design Arenaunit: eloobserved Jul 23, 202610 modelssource ↗

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

websitecodecategoriesdatavizuicomponentgamedev3dsvgasciiartwebappsfullstack+15 more
#ModelScore
1Claude Opus 4.7
1334
2Claude Opus 4.8
1294
3GLM-5.1
1245
4Gemini 3.5 Flash
1244
5Kimi K2.6
1187
6GLM-5V-Turbo
1171
7GPT-5.5
1150
8Gemini 3.1 Pro Preview
1112
9Gemini 3 Flash Preview
1073
10Grok 4.3
1072

Dot = elo on a 1058–1348 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.