llm.ing

Design Arena — androidnative

mirrorrun by Design Arenaunit: elo29 modelssource ↗

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

websitecodecategoriesdatavizuicomponentgamedev3dsvgasciiartwebappsfullstack+15 more
#ModelScoreObserved
1Claude Opus 4.8
1337
Jul 24, 2026
2Claude Fable 5
1319
Jul 24, 2026
3Kimi K2.6
1293
Jul 24, 2026
4Grok 4.5
1276
Jul 24, 2026
5GLM-5V-Turbo
1267
Jul 23, 2026
6Claude Opus 4.7
1262
Jul 24, 2026
7Kimi K2.7 Code
1261
Jul 24, 2026
8Claude Sonnet 5
1253
Jul 24, 2026
9GPT-5.5
1239
Jul 24, 2026
10GLM-5.2
1234
Jul 24, 2026
11GLM-5.1
1233
Jul 24, 2026
12GLM-5
1223
Jul 24, 2026
13Gemini 3.5 Flash
1220
Jul 24, 2026
14Claude Sonnet 4.6
1208
Jul 24, 2026
15Claude Opus 4.6
1191
Jul 24, 2026
16Qwen3.7 Max
1180
Jul 23, 2026
17GPT-5.2 Codex
1176
Jul 23, 2026
18Claude Opus 4.5 (latest)
1164
Jul 24, 2026
19Grok 4.20 (Non-Reasoning)
1153
Jul 24, 2026
20MiniMax-M3
1109
Jul 24, 2026
21Kimi K2.5
1107
Jul 24, 2026
22GLM-4.7
1101
Jul 24, 2026
23GPT-5.3 Codex
1084
Jul 23, 2026
24GLM-4.6
1072
Jul 24, 2026
25Grok 4.3
1056
Jul 24, 2026
26GPT-5.2
1047
Jul 24, 2026
27Gemini 3 Flash Preview
1037
Jul 24, 2026
28Gemini 3.1 Pro Preview
1033
Jul 24, 2026
29GPT-5.4
1015
Jul 24, 2026

Dot = elo on a 998–1354 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.