llm.ing

Design Arena — agenticslides(python-pptx)

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
1331
2Claude Opus 4.8
1310
3Gemini 3.5 Flash
1242
4GLM-5.1
1240
5Kimi K2.6
1186
6GLM-5V-Turbo
1183
7GPT-5.5
1155
8Gemini 3.1 Pro Preview
1107
9Gemini 3 Flash Preview
1075
10Grok 4.3
1068

Dot = elo on a 1054–1345 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.