BENCHMARKS · AI-NATIVE DESIGN · THE LIVE BOARD, MIRRORED

The design leaderboard.
Mirrored, and kept live.

Two views of the same live board: the full Design Arena frame - every arena and cut, both metrics, all three model filters - and the complete 211-model mirror below it. Every number is read off the live board on the date stamped and refreshed hourly - follow the link and check it yourself.

Top leaderboards ↓Full mirror ↓

OPEN AI-NATIVE DESIGN SYSTEM07 / Evaluate

Top leaderboards

The Design Arena frame, mirrored: every arena and cut on the board - code, image, video, audio, slides, SVG and ASCII - with the Elo / win-rate toggle and the All Models / Open Weights / Routers filter, all working off live board data read 18 Sep 2026. Cuts with no public data feed link straight to the live board. Refreshes on the sweep.

Overall · all cuts

Expand ↗

DesignArena · live board data · mirrored 18 sep 2026

64.6%

Kimi K3

64.6%

v0-1.5-lg

64.4%

Gemini 3 Pro Preview

64.3%

Claude Sonnet 4.6

63.3%

Kimi K3

62.9%

Qwen3.8 Max

62%

GPT-6 Astra (xhigh)

61.5%

Grok 4.5

61.2%

Claude Opus 4.6

61.2%

Qwen3 Coder 480B A35B Instruct

61.1%

GLM 5.2

60.9%

GPT-6 Astra (xhigh)

60.8%

Muse Spark

60.6%

Claude Opus 4.6

60.3%

Claude Opus 4.6 (Thinking)

60.2%

Claude Opus 4.5

// win rate · higher is better · top 16 of 213 ranked · colored = top 5 · mirrored from designarena.ai · checked 18 sep 2026

Design Arena

All categories · the full board, mirrored · The full board, mirrored: every model in the all-categories crowd-voted ranking - 1,550,140 votes counted by the board. Elo with standard error, win rate and battles, read off the live board data on the stamped date. The board moves continuously; this mirror refreshes on the site sweep.

#	MODEL	ELO	±	WIN RATE	BATTLES
01	Kimi K3	1388	4.7	64.7%	6,197
02	Muse Spark 1.3 Max	1376	4.4	60.5%	7,175
03	GPT 6 Astra	1369	5.0	62.1%	5,299
04	Muse Spark 1.3	1366	3.6	59.3%	10,685
05	GPT 5.6 Sol Xhigh	1352	4.3	58.2%	7,117
06	Claude Fable 5.1	1345	4.5	59.1%	6,432
07	DeepSeek V4.1 Flash	1344	4.4	56.4%	7,010
08	Claude Fable 5.1	1343	19.1	59.6%	361
09	Claude Opus 5	1338	3.3	58.3%	12,844
10	GPT 5.6 Sol	1334	2.6	57.5%	21,070
11	Kimi K3	1333	11.3	63.6%	1,107
12	Cortado	1331	3.3	58.1%	12,505
13	Claude Opus 5	1329	13.9	62%	726
14	Muse Spark 1.2	1327	2.9	55%	16,475
15	Claude Fable 5	1325	2.9	58.1%	16,702
16	Gemini 3.8 Flash	1323	3.3	54%	13,128
17	Gemini 3.7 Flash	1321	2.0	56.6%	43,316
18	Qwen3.8 Max	1317	10.6	62.5%	1,230
19	GLM 5.2	1310	2.1	53.9%	33,204
20	Muse Spark 1.3	1310	6.6	56.1%	3,844
21	Grogu	1309	2.3	53%	29,583
22	Qwen3.8 Max Preview	1309	4.5	54.1%	6,242
23	Gemini 3.6 Flash	1305	2.4	53.7%	25,173
24	Claude Opus 4.6	1304	2.3	61.2%	27,390
25	Claude Opus 4.6 Thinking	1300	2.5	60.3%	23,804
26	Notebook	1299	2.6	50.3%	20,788
27	Claude Sonnet 5	1292	2.8	53.8%	18,167
28	Yoda	1292	2.2	51.3%	30,388
29	Claude Sonnet 4.6	1291	2.3	59.4%	26,300
30	Evergreen	1291	3.0	52.1%	15,472
31	Deca 2.6 Ultra	1290	7.5	49.6%	2,209
32	Qwen3.7 Max	1290	2.2	55%	29,920
33	Kimi K2.6	1287	2.1	54.8%	34,014
34	MiMo V2.5 Pro	1287	2.5	53.3%	22,406
35	Deca 2.5 Mini	1286	5.5	52.2%	4,198
36	Claude Fable 5	1283	9.0	59.4%	1,733
37	Coconut	1282	2.4	53.8%	24,305
38	Qwen3.7 Plus	1281	3.3	50.6%	12,401
39	GLM 5 Turbo	1280	2.0	54.8%	36,972
40	Deca 2.5 Ultra 2	1278	5.7	51%	3,923
41	Gemini 3.5 Flash	1278	2.2	54.7%	30,071
42	GPT 5.6 Sol Xhigh	1278	9.5	56.8%	1,587
43	Grogu	1277	6.1	55.6%	4,284
44	MiMo V2.5	1275	2.0	54.2%	39,751
45	GPT 5.5	1274	2.3	54.2%	27,474
46	Kimi K2.7 Code	1271	2.8	51.2%	17,856
47	Radon	1267	5.6	60.8%	4,250
48	Claude Opus 4.8	1266	2.3	53.3%	26,261
49	MiniMax M3	1265	2.5	52%	23,222
50	Yoda	1263	7.9	62%	2,513
51	Claude Sonnet 5	1262	7.2	55.3%	2,766
52	GLM 5	1261	1.8	55.5%	45,343
53	Claude Opus 4.8	1261	7.9	57.1%	2,272
54	Riftrunner	1259	2.4	64.4%	25,342
55	Gemini 3.8 Flash	1259	16.4	48.1%	482
56	Claude Opus 4.5	1257	2.0	59.5%	35,203
57	GPT 5.6 Terra	1256	1.7	48.7%	67,876
58	DeepSeek V4 Pro	1255	2.3	51.9%	27,274
59	GLM 5.2	1252	7.6	61.2%	2,572
60	Kimi K2.5	1251	1.9	54.1%	41,610
61	Qwen3.6 Plus	1250	2.1	51.1%	33,380
62	Deepocto	1249	1.9	51.7%	40,421
63	GPT 5.6 Luna	1247	3.0	47.4%	14,717
64	Nex N2 Pro	1247	3.2	48%	13,041
65	Gemini 3.1 Pro Preview	1246	1.8	54.5%	46,903
66	DeepSeek V4 Flash 0731	1245	3.6	46.5%	10,435
67	GLM 5v Turbo	1244	1.9	50.6%	41,418
68	Hy3 Direct	1244	2.5	46.9%	22,527
69	Grok 4 20 Beta Reasoning	1236	2.0	51.8%	36,709
70	GLM 4.7	1232	1.8	54.8%	46,837
71	Claude Opus 4.6	1231	7.9	61.2%	2,310
72	Lavender	1227	3.7	52.5%	9,608
73	Galapagos	1225	2.7	52.7%	18,942
74	Claude Sonnet 4.6	1225	8.6	64.3%	2,030
75	Gemini 3.7 Flash	1224	16.9	45.7%	442
76	Muse Spark 1.2	1224	9.1	49.3%	1,628
77	Camellia	1223	6.3	43.4%	3,184
78	Deepmolt	1222	3.4	56.8%	11,506
79	Grok 4 20 Beta Non Reasoning	1221	1.9	49.6%	39,854
80	DeepSeek V4 Flash	1220	2.2	48.9%	28,076
81	Evergreen	1218	10.5	47.5%	1,218
82	Inkling	1215	3.8	42.1%	9,373
83	MiniMax M3	1211	7.5	48.9%	2,542
84	Gemini 3.5 Flash	1210	6.3	57.1%	3,834
85	MiniMax M2.1	1205	2.6	55.3%	20,805
86	Gemini 3 Flash Preview	1204	5.4	57.6%	4,414
87	Qwen3.7 Max	1204	10.6	49.8%	1,174
88	Grok 4.3	1201	2.1	44.9%	33,469
89	Claude Sonnet 4.5	1198	1.8	51.5%	45,427
90	Claude Sonnet 4.5 Thinking	1197	1.8	51.3%	44,186
91	Solar Pro4 Direct	1197	2.9	38.9%	17,826
92	GPT 5.4 Low	1195	2.5	49.2%	21,139
93	Qwen3.5 397b A17b	1195	4.0	52.6%	8,129
94	Step 3.7 Flash	1195	2.5	43.5%	23,577
95	Gemini 3.6 Flash	1195	7.7	45.7%	2,410
96	GPT 5.4 None	1193	2.4	49.2%	23,602
97	GLM 4.7 Flash	1192	3.3	53.1%	11,706
98	Claude 3.7 Sonnet	1191	3.0	57.6%	15,259
99	Hy3	1191	5.7	41.1%	4,063
100	DeepSeek V3p1 Thinking	1190	2.9	56%	16,257
101	Claude Opus 4.1 20250805 Thinking	1186	2.9	55.8%	15,715
102	DeepSeek V3p2 Exp	1186	2.6	54.2%	19,487
103	GPT 5.1 High	1186	2.9	53%	16,116
104	Kimi K2.7 Code	1186	7.4	51.8%	2,577
105	Rockhopper	1185	2.2	49.6%	27,932
106	Mumble	1184	2.2	49.8%	28,873
107	GPT 5 High	1183	3.1	54.5%	13,485
108	Qwen3.5 Plus 02 15	1182	2.7	48.5%	18,498
109	Macaroni	1181	2.3	50.3%	25,456
110	Claude Opus 4.1 20250805	1180	1.9	50.1%	39,734
111	GLM 4.5	1180	2.6	54.4%	19,653
112	GLM 4.6	1180	2.8	54.3%	17,086
113	GPT 5	1180	2.1	51.8%	32,965
114	DeepSeek V3p2	1179	2.2	49.3%	29,122
115	GLM 5.1	1177	6.3	53.5%	3,747
116	Claude Opus 4	1176	2.8	55.6%	16,668
117	GPT 5.1 Medium	1176	2.5	51.9%	21,347
118	GPT 5.1 Low	1171	2.5	51.5%	22,192
119	MiMo V2 Flash	1171	2.0	49.4%	35,226
120	GPT 5.6 Sol	1170	9.2	49.8%	1,612
121	Claude Opus 4.5 20251101	1168	10.1	60.2%	1,395
122	Kimi K2.6	1168	5.5	51%	5,241
123	Gemini 2.5 Pro	1167	4.3	57.5%	7,220
124	GPT 5.1 Codex	1167	7.9	54.6%	2,005
125	GPT 5.1 None	1166	2.5	50.7%	22,295
126	Honeycomb	1165	5.5	50.5%	4,150
127	GLM 5v Turbo	1161	6.4	52%	3,939
128	GPT 5.3 Codex	1159	2.9	47.2%	15,673
129	Qwen3 Coder 480b A35b Instruct	1158	8.3	61.2%	1,955
130	V0.1.5 Lg	1158	10.9	64.6%	1,192
131	Mistral Large 2512	1157	2.2	47.6%	30,428
132	Claude Sonnet 4	1156	2.8	53.4%	17,542
133	Nemotron 3 Ultra 550b A55b	1154	2.7	36.3%	19,469
134	DeepSeek Reasoner R1	1153	2.7	52.6%	17,951
135	Agi 01 Swift	1152	10.3	38.1%	1,248
136	GLM 4.5 Air	1152	2.8	51.4%	17,253
137	Claude Sonnet 4 Thinking	1151	2.9	51.5%	16,246
138	MiniMax M2 Stable	1150	3.5	48.1%	10,828
139	Aescoder 4b	1140	1.9	45.9%	39,833
140	Mistral Medium 3.5	1136	3.5	41.1%	10,894
141	Mistral Medium 2508	1135	2.2	45.1%	27,823
142	GLM 5	1134	5.7	51.7%	4,762
143	Trinity Large Thinking	1132	3.3	40.1%	12,837
144	Claude Haiku 4.5	1130	2.0	44.9%	35,538
145	GPT 5 Mini	1130	2.0	43.5%	37,081
146	DeepSeek V3p1	1127	2.6	47.8%	20,334
147	Qwen3 Max	1125	2.1	44%	33,366
148	Kimi K2.5	1125	8.8	54.5%	1,787
149	DeepSeek Chat	1123	2.6	48.4%	19,271
150	Intellect 3	1121	2.1	42.6%	31,346
151	Gemini 2.5 Flash Preview 09 2025	1119	2.6	46.9%	19,295
152	Grok 4 Fast Non Reasoning	1115	2.0	43%	36,646
153	Kimi K2 0905 Preview	1113	9.1	48.5%	1,504
154	GPT 5.1 Codex Mini	1109	2.1	41.5%	33,827
155	Grok 4.1 Fast Non Reasoning	1107	2.1	42.1%	33,598
156	Grok 4.1 Fast Reasoning	1102	2.2	41.2%	31,132
157	GPT 5.5	1101	6.5	43%	3,537
158	GPT 5 Nano	1099	4.4	48%	6,703
159	Kimi K2 Turbo Preview	1098	7.8	46.9%	2,094
160	Gemini 2.5 Flash Lite Preview 09 2025	1096	4.3	47%	6,850
161	Gemini 3.1 Flash Lite Preview	1087	2.5	36.4%	23,527
162	Gemini 3 Pro Preview	1086	8.0	48.7%	2,259
163	Mistral Medium 2505	1084	4.5	48.1%	6,390
164	Gemini 3.1 Pro Preview	1081	5.8	41.6%	4,614
165	Ministral 14b 2512	1080	7.3	44%	2,380
166	Gemini 3 Flash Preview	1075	7.8	47.4%	2,358
167	Gemini 2.5 Flash	1074	4.3	45.4%	6,971
168	GPT 5.6 Terra	1074	8.3	31.5%	2,388
169	V0.1.5 Md	1071	3.5	40.2%	11,081
170	Grok 3	1068	2.3	40.1%	26,865
171	Ministral 8b 2512	1068	7.3	42.9%	2,427
172	Grok 4 Fast Reasoning	1065	2.0	36.7%	37,369
173	Grok 4.20 0309 Reasoning	1065	7.8	41%	2,424
174	Claude Sonnet 4.5 20250929	1062	14.3	43.7%	629
175	GLM 4.7	1061	9.2	45.1%	1,638
176	GPT 5.2	1054	8.2	44.3%	2,165
177	Qwen3 235b A22b Instruct 2507	1053	4.4	42.6%	6,939
178	Kimi K2 0711 Preview	1048	9.9	51.7%	1,352
179	Magistral Medium 2509	1048	4.8	40.4%	5,854
180	Qwen3 235b A22b Thinking 2507	1048	4.6	40.8%	6,175
181	GLM 4.6	1041	8.7	42.3%	1,885
182	GPT 4.1	1040	8.8	50.9%	1,747
183	GPT O3	1034	10.0	51.9%	1,365
184	Grok 4	1031	2.5	35%	24,016
185	GPT 5.4	1028	8.3	40.8%	2,151
186	Devstral Medium 2507	1027	4.4	41.2%	7,153
187	Mercury 2	1026	3.3	21.2%	18,418
188	Ministral 3b 2512	1025	6.9	37.3%	2,852
189	Codestral 2508	1022	4.5	38.5%	6,734
190	Grok Code Fast 1	1021	8.5	39.7%	2,004
191	Qwen3 235b A22b	1017	5.2	38.2%	5,158
192	Grok 4.3	1017	7.3	28.5%	3,233
193	Grok Code Fast 1	1013	5.6	36.8%	4,299
194	GPT 4.1 Mini	1008	9.3	47.5%	1,566
195	GPT 5.2 Codex	1004	8.4	37.2%	2,157
196	Magistral Small 2509	1001	4.7	34.6%	6,447
197	GPT 5.3 Codex	994	11.3	36.5%	1,109
198	GPT O4 Mini	991	8.3	46.4%	2,011
199	Olmo 3.1 32b Think	990	3.2	27.1%	16,164
200	GPT 4.1 Nano	978	8.6	47.3%	1,901
201	GPT Oss 120b	978	5.2	33.4%	5,272
202	Qwen3 30b A3b	957	7.4	37.5%	2,568
203	Grok 3 Mini	945	4.5	32.4%	7,621
204	Llama 3.1 Nemotron Ultra 253b	944	7.1	27.2%	3,173
205	Mistral Small 2506	922	10.6	39.8%	1,243
206	Llama 4 Maverick	894	9.4	35.8%	1,678
207	Mistral Large 2411	878	10.7	33.3%	1,317
208	GPT 4o	875	9.3	34.8%	1,780
209	Codestral 2 2501	848	10.6	30.2%	1,444
210	Devstral Small 2507	822	11.5	27.7%	1,250
211	Llama 4 Scout	804	11.6	26.6%	1,275

// elo · higher is better · shading = top 5 per column · mirrored from designarena.ai/leaderboard · checked 13 sep 2026