llm.ing

Design Arena — godotgamedev

mirrorrun by Design Arenaunit: elo29 modelssource ↗

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

websitecodecategoriesdatavizuicomponentgamedev3dsvgasciiartwebappsfullstack+15 more
#ModelScoreObserved
1Claude Fable 5
1300
Jul 24, 2026
2Grok 4.5
1296
Jul 24, 2026
3Claude Opus 4.7
1270
Jul 23, 2026
4Qwen3.7 Max
1249
Jul 23, 2026
5Claude Sonnet 5
1239
Jul 23, 2026
6Gemini 3.1 Pro Preview
1236
Jul 23, 2026
7Claude Opus 4.8
1233
Jul 24, 2026
8Claude Sonnet 4.6
1226
Jul 23, 2026
9Kimi K2.5
1219
Jul 23, 2026
10Kimi K2.7 Code
1209
Jul 24, 2026
11Kimi K3
1199
Jul 23, 2026
12GPT-5.5
1187
Jul 23, 2026
13GLM-4.6
1180
Jul 23, 2026
14Gemini 3.5 Flash
1172
Jul 24, 2026
15Kimi K2.6
1168
Jul 23, 2026
16Gemini 3 Flash Preview
1161
Jul 23, 2026
17GLM-5
1144
Jul 23, 2026
18GPT-5.2 Codex
1142
Jul 23, 2026
18GPT-5.2
1142
Jul 23, 2026
18GLM-5.2
1142
Jul 23, 2026
21Muse Spark 1.1
1136
Jul 24, 2026
22GPT-5.4
1135
Jul 23, 2026
23GPT-5.3 Codex
1124
Jul 23, 2026
24GLM-5.1
1104
Jul 23, 2026
25Grok 4.20 (Non-Reasoning)
1089
Jul 23, 2026
26DeepSeek V4 Pro
1059
Jul 23, 2026
27GLM-4.7
1058
Jul 23, 2026
28GLM-5V-Turbo
1043
Jul 23, 2026
29Grok 4.3
1035
Jul 23, 2026

Dot = elo on a 1021–1314 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.