llm.ing

Design Arena — python-pptxslides

mirrorrun by Design Arenaunit: elo18 modelssource ↗

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

websitecodecategoriesdatavizuicomponentgamedev3dsvgasciiartwebappsfullstack+15 more
#ModelScoreObserved
1Claude Opus 4.7
1336
Jul 23, 2026
2Claude Fable 5
1307
Jul 23, 2026
3Claude Opus 4.8
1298
Jul 23, 2026
4Kimi K3
1280
Jul 24, 2026
5GLM-5.1
1258
Jul 23, 2026
6Claude Sonnet 5
1251
Jul 24, 2026
7Gemini 3.5 Flash
1247
Jul 23, 2026
8Grok 4.5
1233
Jul 24, 2026
9Qwen3.7 Max
1229
Jul 23, 2026
10GLM-5.2
1188
Jul 24, 2026
11Kimi K2.6
1180
Jul 23, 2026
12Muse Spark 1.1
1175
Jul 23, 2026
13GLM-5V-Turbo
1165
Jul 23, 2026
14GPT-5.5
1152
Jul 23, 2026
15Kimi K2.7 Code
1151
Jul 23, 2026
16Gemini 3.1 Pro Preview
1109
Jul 23, 2026
17Grok 4.3
1072
Jul 23, 2026
18Gemini 3 Flash Preview
1021
Jul 24, 2026

Dot = elo on a 1005–1352 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.