Design Arena Leaderboard

312 models across 17 categories · 5.3M+ votes · Updated Jul 25, 2026

Crowdsourced design benchmark from designarena.ai — ELO ratings based on real user preferences

Game Dev (Agentic)

20 models · 13,770 votes

#	Model	Provider	ELO ▼	Win%	Battles	CI 95%	Avg Time
1	Claude Fable 5	anthropic	1297	65.0%	816	±0.0	nullms
2	Claude Opus 4.7	anthropic	1254	61.5%	1,521	±0.0	nullms
3	Claude Opus 4.8	anthropic	1254	61.1%	835	±0.0	nullms
4	Claude Sonnet 5	anthropic	1231	55.0%	742	±0.0	nullms
5	MiniMax M3	minimax	1213	54.9%	215	±0.0	nullms
6	Grok 4.5	xai	1210	52.7%	712	±0.0	nullms
7	GPT-5.6 Sol	openai	1207	52.2%	1,116	±0.0	nullms
8	Claude Sonnet 4.6	anthropic	1205	55.1%	214	±0.0	nullms
9	Muse Spark 1.1	meta	1189	47.9%	2,596	±0.0	nullms
10	Gemini 3.5 Flash	google	1186	54.0%	3,912	±0.0	nullms
11	GLM 5.2	zhipu	1184	48.4%	341	±0.0	nullms
12	GPT-5.5	openai	1184	51.9%	2,201	±0.0	nullms
13	GLM 5.1	zhipu	1175	50.8%	1,389	±0.0	nullms
14	Qwen3.7 Max	alibaba	1175	48.3%	654	±0.0	nullms
15	Kimi K2.6	moonshot	1152	48.4%	2,260	±0.0	nullms
16	Kimi K2.7 Code	moonshot	1137	42.9%	753	±0.0	nullms
17	Gemini 3.1 Pro Preview	google	1121	43.7%	1,848	±0.0	nullms
18	GLM 5V Turbo	zhipu	1108	41.8%	1,989	±0.0	nullms
19	GPT-5.6 Terra	openai	1086	37.4%	278	±0.0	nullms
20	Grok 4.3	xai	1011	28.3%	1,756	±0.0	nullms