llm.ing

Design Arena — htmlslides

mirrorrun by Design Arenaunit: elo25 modelssource ↗

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

websitecodecategoriesgamedev3duicomponentdatavizsvgasciiartmobileappswebapps+19 more
#ModelScoreObserved
1Claude Fable 5.1
1393
Nov 19, 2026
2Claude Opus 4.7
1217
Sep 6, 2026
3GLM-5.3
1189
Sep 11, 2026
4Grok 4.20 (Non-Reasoning)
1180
Aug 18, 2026
5Claude Fable 5
1176
Nov 19, 2026
6Kimi K3
1155
Nov 19, 2026
7Grok 4.6
1148
Nov 19, 2026
8Claude Opus 4.8
1121
Nov 19, 2026
9Claude Sonnet 5
1116
Nov 19, 2026
10Grok 4.5
1102
Nov 19, 2026
11Kimi K2.7 Code
1101
Nov 19, 2026
12Muse Spark 1.1
1094
Nov 19, 2026
13Kimi K2.6
1084
Nov 19, 2026
14MiniMax-M3
1080
Nov 19, 2026
15GLM-5.2
1073
Nov 19, 2026
16GLM-5.1
1070
Nov 19, 2026
17Qwen3.7 Max
1056
Nov 19, 2026
18Gemini 3.1 Pro Preview
1052
Nov 19, 2026
19Gemini 3.6 Flash
1047
Nov 19, 2026
20GLM-5
1046
Nov 19, 2026
21Gemini 3.5 Flash
1042
Nov 19, 2026
22Muse Spark 1.2
1031
Nov 19, 2026
23GLM-5V-Turbo
1020
Nov 19, 2026
24GPT-5.5
966
Nov 19, 2026
25Grok 4.3
909
Nov 19, 2026

Dot = elo on a 884–1418 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.