Skip to main content

Benchmark profile

Design Arena Website Elo (Design Arena Website)

A display-only Design Arena website-generation Elo score surfaced on OpenRouter model benchmark pages.

Data verified

Benchmark score on Design Arena Website — July 29, 2026

BenchLM mirrors the published score view for Design Arena Website. Kimi K3 leads the public snapshot at 1377 , followed by Claude Opus 5 (1341) and GLM-5.2 (1341). BenchLM does not use these results to rank models overall.

89 modelsMultimodal & GroundedCurrentDisplay onlyUpdated July 29, 2026

Benchmark score table (89 models)

Score
1
Kimi K3Moonshot AI · Closed
1377
2
Claude Opus 5Anthropic · Closed
1341
3
GLM-5.2Z.AI · Open weight
1341
4
Claude Fable 5Anthropic · Closed
1324
5
Grok 4.5xAI · Closed
1321
6
Claude Opus 4.7 (Adaptive)Anthropic · Closed
1320
7
Claude Opus 4.7Anthropic · Closed
1320
8
Claude Opus 4.6Anthropic · Closed
1319
9
Gemini 3.6 FlashGoogle · Closed
1319
10
Claude Opus 4.6 (Adaptive)Anthropic · Closed
1319
11
Claude Sonnet 5Anthropic · Closed
1314
12
Claude Sonnet 4.6Anthropic · Closed
1310
13
MiMo-V2.5-ProXiaomi · Closed
1309
14
Qwen3.7 MaxAlibaba · Closed
1303
15
Kimi K2.6Moonshot AI · Open weight
1302
16
Kimi K2.7 CodeMoonshot AI · Open weight
1300
17
GLM-5.1Z.AI · Open weight
1299
18
Qwen3.7 PlusAlibaba · Closed
1297
19
GLM-5-TurboZ.AI · Closed
1296
20
MiMo-V2.5Xiaomi · Closed
1295
21
Muse Spark 1.1Meta · Closed
1291
22
MiniMax M3MiniMax · Open weight
1286
23
Gemini 3.5 FlashGoogle · Closed
1283
24
GPT-5.5OpenAI · Closed
1280
25
Gemini 3.1 ProGoogle · Closed
1276
26
Kimi K2.5Moonshot AI · Open weight
1275
27
Kimi K2.5 (Reasoning)Moonshot AI · Closed
1275
28
GLM-5Z.AI · Open weight
1274
29
GLM-5 (Reasoning)Z.AI · Open weight
1274
30
Claude Opus 4.5Anthropic · Closed
1272
31
Claude Opus 4.5 ThinkingAnthropic · Closed
1272
32
MiniMax M2.7MiniMax · Open weight
1271
33
Claude Opus 4.8Anthropic · Closed
1268
34
Qwen3.6 PlusAlibaba · Closed
1265
35
DeepSeek V4 Pro (Max)DeepSeek · Open weight
1260
36
DeepSeek V4 Pro (High)DeepSeek · Open weight
1260
37
DeepSeek V4 ProDeepSeek · Open weight
1260
38
GLM-5V-TurboZ.AI · Closed
1255
39
Grok 4.20xAI · Closed
1252
40
GLM-4.7Z.AI · Open weight
1251
41
MiniMax M2.5MiniMax · Closed
1247
42
GPT-5.4OpenAI · Closed
1245
43
DeepSeek V4 Flash (Max)DeepSeek · Open weight
1233
44
DeepSeek V4 Flash (High)DeepSeek · Open weight
1233
45
DeepSeek V4 FlashDeepSeek · Open weight
1233
46
Gemini 3 FlashGoogle · Closed
1221
47
GPT-5.2OpenAI · Closed
1219
48
Grok 4.3xAI · Closed
1219
49
GLM-4.7-FlashZ.AI · Open weight
1219
50
Claude Sonnet 4.5Anthropic · Closed
1215
51
Claude Sonnet 4.5 ThinkingAnthropic · Closed
1215
52
GPT-5.1OpenAI · Closed
1212
53
Step 3.7 FlashStepFun · Open weight
1212
54
GPT-5 (high)OpenAI · Closed
1209
55
GPT-5 (medium)OpenAI · Closed
1209
56
Hy3Tencent · Open weight
1206
57
Claude 4.1 OpusAnthropic · Closed
1202
58
DeepSeek V3.2DeepSeek · Open weight
1200
59
DeepSeek V3.2 (Thinking)DeepSeek · Open weight
1200
60
GLM-4.5Z.AI · Closed
1195
61
Gemini 2.5 ProGoogle · Closed
1192
62
GPT-5.3 CodexOpenAI · Closed
1187
63
GPT-5.1-CodexOpenAI · Closed
1186
64
GLM-4.5-AirZ.AI · Closed
1171
65
Claude 4 SonnetAnthropic · Closed
1170
66
Trinity-Large-PreviewArcee AI · Open weight
1162
67
Trinity-Large-ThinkingArcee AI · Open weight
1162
68
GPT-5 miniOpenAI · Closed
1150
69
Claude Haiku 4.5Anthropic · Closed
1147
70
DeepSeek V3.1 (Reasoning)DeepSeek · Open weight
1147
71
DeepSeek V3.1DeepSeek · Open weight
1147
72
Claude Haiku 4.5 ThinkingAnthropic · Closed
1147
73
DeepSeek V3DeepSeek · Open weight
1145
74
Qwen3 MaxAlibaba · Closed
1144
75
Gemini 2.5 FlashGoogle · Closed
1140
76
Nemotron 3 UltraNVIDIA · Open weight
1127
77
GPT-5 nanoOpenAI · Closed
1126
78
Mistral Medium 3Mistral · Closed
1104
79
Kimi K2Moonshot AI · Closed
1075
80
GPT-4.1OpenAI · Closed
1063
81
o3OpenAI · Closed
1061
82
GPT-4.1 miniOpenAI · Closed
1022
83
Mercury 2Inception · Closed
1010
84
GPT-4.1 nanoOpenAI · Closed
998
85
GPT-OSS 120BOpenAI · Open weight
993
86
Llama 4 MaverickMeta · Open weight
896
87
GPT-OSS 20BOpenAI · Open weight
877
88
GPT-4oOpenAI · Closed
856
89
Llama 4 ScoutMeta · Open weight
775

The published Design Arena Website snapshot places Kimi K3 first at 1377. The third row is 36 score units behind. The broader top-10 range is 58 score units, so the table still separates the published systems.

89 models have been evaluated on Design Arena Website. The benchmark falls in the Multimodal & Grounded category. This category carries a 12% weight in BenchLM.ai's overall scoring system. Design Arena Website is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About Design Arena Website

Year

2026

Tasks

Website generation comparisons

Format

Elo

Difficulty

Design and website generation

OpenRouter's Grok 4.3 benchmark page reports Website at 1294 Elo, 56.5% win rate, 166.3s average generation time, and Top 13%. BenchLM stores the Elo as the benchmark value and keeps the supporting fields in the description.

BenchLM freshness & provenance

Version

Design Arena Website 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

FAQ

What does Design Arena Website measure?

A display-only Design Arena website-generation Elo score surfaced on OpenRouter model benchmark pages.

Which model scores highest on Design Arena Website?

Kimi K3 by Moonshot AI currently leads with a score of 1377 on Design Arena Website.

How many models are evaluated on Design Arena Website?

89 AI models have been evaluated on Design Arena Website on BenchLM.

Last updated: July 29, 2026 · BenchLM version Design Arena Website 2026

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.