llm.ing

Design Arena — python-pptxslides

mirrorrun by Design Arenaunit: elo27 modelssource ↗

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

websitecodecategoriesdatavizuicomponentgamedev3dsvgasciiartfullstackmobileapps+15 more
#ModelScoreObserved
1Claude Opus 4.7
1336
Jul 23, 2026
2Claude Fable 5.1
1328
Sep 11, 2026
3Claude Opus 4.8
1298
Jul 23, 2026
4Claude Fable 5
1294
Sep 20, 2026
5Kimi K3
1273
Sep 11, 2026
6Claude Opus 5
1265
Sep 15, 2026
7GLM-5.1
1258
Jul 23, 2026
7GLM-5.3
1258
Sep 18, 2026
9Gemini 3.5 Flash
1247
Jul 23, 2026
10Muse Spark 1.3
1245
Sep 11, 2026
11Grok 4.6
1242
Sep 20, 2026
12Grok 4.5
1230
Sep 11, 2026
13Claude Sonnet 5
1216
Sep 12, 2026
14MiniMax-M3
1207
Sep 13, 2026
15Muse Spark 1.2
1198
Sep 12, 2026
16Qwen3.7 Max
1194
Sep 11, 2026
16GLM-5.2
1194
Sep 12, 2026
18Kimi K2.6
1180
Jul 23, 2026
19Gemini 3.8 Flash
1176
Sep 13, 2026
20GLM-5V-Turbo
1165
Sep 5, 2026
21Muse Spark 1.1
1156
Sep 12, 2026
22Kimi K2.7 Code
1154
Sep 12, 2026
23GPT-5.5
1152
Jul 23, 2026
24Gemini 3.6 Flash
1146
Sep 13, 2026
25Gemini 3.1 Pro Preview
1109
Jul 23, 2026
26Grok 4.3
1072
Jul 23, 2026
27Gemini 3 Flash Preview
1036
Aug 18, 2026

Dot = elo on a 1021–1351 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.