Benchmark Almanac

DesignArena SVG

Arrow 2 Telos 1,502 tops its leaderboard GPT-6 Astra 1,423 leads the models we track
Not in model cards

Models write SVG code for a design request such as an icon, logo, badge or illustration. Users compare the rendered graphics and vote.

One of 40 DesignArena boards. DesignArena ranks entries with Bradley-Terry ratings from user votes.

Kind
Web and app front ends
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
Strongest model tested
GPT-6 Astra #3 on the capability index
Last model added
October 2026
Scores refreshed
November 12th, 2026
119
Entries on its own leaderboard, 103 models we track

Arrow 2 Telos tops the leaderboard with 1,502

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is GPT-6 Astra, at 1,423.

119 entries. Scroll sideways for the rest. Bars start at 800.

  1. 1,502 Arrow 2 Telos
  2. 1,423 GPT-6 Astra
  3. 1,369 Arrow 2
  4. 1,335 Gemini 4 Argon
  5. 1,327 Kimi K3.1 Pro
  6. 1,321 Claude Opus 5.5
  7. 1,316 Muse Spark 1.3
  8. 1,307 Claude Opus 5
  9. 1,305 GPT-5.6 Sol
  10. 1,303 GPT-6 Sol
  11. 1,303 Claude Sonnet 5.5
  12. 1,298 Kimi K3
  13. 1,294 Claude Fable 5
  14. 1,294 Claude Fable 5.1
  15. 1,294 GPT-6.1 Terra
  16. 1,293 MiMo-V2.6 Pro
  17. 1,291 Gemini 3.8 Flash
  18. 1,282 MiniMax-M3.1-Flash
  19. 1,280 Gemini 3.7 Flash
  20. 1,276 Gemini 3.1 Pro
  21. 1,274 Muse Spark 1.2
  22. 1,272 GLM-5.3-Flash
  23. 1,266 Arrow 1.1 Max
  24. 1,266 Ling-3.1-flash
  25. 1,265 Arrow-1.0
  26. 1,248 GLM-5.3
  27. 1,245 Gemini 3.5 Flash
  28. 1,244 DeepSeek V4.1 Flash
  29. 1,238 GPT-6 Luna
  30. 1,235 Gemini 3 Pro
  31. 1,226 GPT-5.5
  32. 1,225 Muse Spark 1.1
  33. 1,224 Mistral Large 4
  34. 1,221 Grok 4.6
  35. 1,218 Claude Opus 4.6
  36. 1,217 Recraft V4.1 Vector Pro
  37. 1,217 Muse Spark
  38. 1,210 Recraft V4 Vector
  39. 1,209 GLM-5.1
  40. 1,209 GPT-5.6 Terra
  41. 1,208 Recraft V4 Vector Pro
  42. 1,207 Qwen3.7-Max
  43. 1,202 Qwen3.7-Plus
  44. 1,201 Grok 4.5
  45. 1,197 GLM-5.2
  46. 1,196 Arrow 1.1
  47. 1,195 GLM-5-Turbo
  48. 1,194 Tencent Hy3
  49. 1,194 Claude Haiku 5.5
  50. 1,187 Nex N2 Pro
  51. 1,187 Claude Sonnet 4.6
  52. 1,187 GPT-5.6 Luna
  53. 1,186 Claude Sonnet 5
  54. 1,185 GPT-5.4
  55. 1,183 MiMo-V2.5-Pro
  56. 1,183 Claude Opus 4.8
  57. 1,179 GPT-5
  58. 1,175 MiMo-V2.5
  59. 1,172 Kimi K2.6
  60. 1,171 Claude Opus 4.5
  61. 1,171 DeepSeek V4 Flash 0731
  62. 1,167 Mercury 3
  63. 1,166 Kimi K2.7 Code
  64. 1,159 GLM-5
  65. 1,154 Grok 4.20 Beta
  66. 1,152 Qwen 3.6 Plus
  67. 1,147 MiniMax-M3
  68. 1,146 DeepSeek-V4-Flash
  69. 1,143 Claude Opus 4.1
  70. 1,140 Kimi K2.5
  71. 1,140 GLM-5V-Turbo
  72. 1,139 GPT-5.1
  73. 1,139 MiniMax-M2.5
  74. 1,133 GLM-4.7
  75. 1,130 Qwen3.5 397B-A17B
  76. 1,129 GPT-5.2
  77. 1,129 DeepSeek-V4-Pro
  78. 1,124 MiniMax-M2.7
  79. 1,122 MiniMax-M2.1
  80. 1,119 GPT-5.3 Codex
  81. 1,118 Claude Opus 4
  82. 1,105 Claude Sonnet 4.5
  83. 1,102 Inkling
  84. 1,099 GLM-4.6
  85. 1,098 Qwen3.5 Plus 02-15
  86. 1,090 MiniMax M2 Stable
  87. 1,086 GLM-4.5
  88. 1,081 GPT-5 mini (Default)
  89. 1,078 Claude 3.7 Sonnet
  90. 1,074 Step 3.7 Flash
  91. 1,072 Grok 4.3
  92. 1,069 Claude Sonnet 4
  93. 1,062 GLM-4.5-Air
  94. 1,061 Nemotron 3 Ultra
  95. 1,055 DeepSeek-V3.1
  96. 1,052 v0-1.5-md
  97. 1,043 Gemini 3.1 Flash-Lite
  98. 1,029 DeepSeek-R1
  99. 1,028 glm-4.7-flash
  100. 1,023 DeepSeek-V3.2
  101. 1,022 DeepSeek-V3.2-Exp
  102. 1,020 MiMo-V2-Flash
  103. 1,016 Claude Haiku 4.5
  104. 1,011 Grok 4
  105. 1,010 Gemini 2.5 Flash
  106. 1,007 Trinity Large Thinking
  107. 1,002 Qwen3-Max
  108. 999 Grok 4.1 Fast
  109. 991 Mercury 2
  110. 983 Mistral Medium 3.1
  111. 983 Mistral Large 3
  112. 982 Prime Intellect: INTELLECT-3
  113. 973 Grok 3
  114. 968 GPT-5.1-Codex-mini
  115. 967 DeepSeek-V3
  116. 966 Grok 4 Fast
  117. 940 Mistral Medium 3.5
  118. 921 AesCoder-4B
  119. 862 Olmo 3.1 32B Think

Bars start at 800.

  1. 1 Arrow 2 Telos 1,502
  2. 2 GPT-6 Astra xhigh 1,423
  3. 3 Arrow 2 1,369
  4. 4 Gemini 4 Argon 1,335
  5. 5 Kimi K3.1 Pro 1,327
  6. 6 Claude Opus 5.5 1,321
  7. 7 Muse Spark 1.3 xhigh, max 1,316
  8. 8 Claude Opus 5 1,307
  9. 9 GPT-5.6 Sol medium, xhigh 1,305
  10. 10 GPT-6 Sol 1,303
  11. 11 Claude Sonnet 5.5 1,303
  12. 12 Kimi K3 1,298
  13. 13 Claude Fable 5 1,294
  14. 14 Claude Fable 5.1 1,294
  15. 15 GPT-6.1 Terra 1,294
  16. 16 MiMo-V2.6 Pro 1,293
  17. 17 Gemini 3.8 Flash 1,291
  18. 18 MiniMax-M3.1-Flash 1,282
  19. 19 Gemini 3.7 Flash 1,280
  20. 20 Gemini 3.1 Pro 1,276
  21. 21 Muse Spark 1.2 1,274
  22. 22 GLM-5.3-Flash 1,272
  23. 23 Arrow 1.1 Max 1,266
  24. 24 Ling-3.1-flash 1,266
  25. 25 Arrow-1.0 1,265
  26. 26 GLM-5.3 1,248
  27. 27 Gemini 3.5 Flash 1,245
  28. 28 DeepSeek V4.1 Flash 1017 1,244
  29. 29 GPT-6 Luna 1,238
  30. 30 Gemini 3 Pro 1,235
  31. 31 GPT-5.5 1,226
  32. 32 Muse Spark 1.1 1,225
  33. 33 Mistral Large 4 1,224
  34. 34 Grok 4.6 1,221
  35. 35 Claude Opus 4.6 thinking 1,218
  36. 36 Recraft V4.1 Vector Pro 1,217
  37. 37 Muse Spark 1,217
  38. 38 Recraft V4 Vector 1,210
  39. 39 GLM-5.1 1,209
  40. 40 GPT-5.6 Terra 1,209
  41. 41 Recraft V4 Vector Pro 1,208
  42. 42 Qwen3.7-Max 1,207
  43. 43 Qwen3.7-Plus 1,202
  44. 44 Grok 4.5 1,201
  45. 45 GLM-5.2 1,197
  46. 46 Arrow 1.1 1,196
  47. 47 GLM-5-Turbo 1,195
  48. 48 Tencent Hy3 1,194
  49. 49 Claude Haiku 5.5 1,194
  50. 50 Nex N2 Pro 1,187
  51. 51 Claude Sonnet 4.6 1,187
  52. 52 GPT-5.6 Luna 1,187
  53. 53 Claude Sonnet 5 1,186
  54. 54 GPT-5.4 medium 1,185
  55. 55 MiMo-V2.5-Pro 1,183
  56. 56 Claude Opus 4.8 1,183
  57. 57 GPT-5 minimal, high 1,179
  58. 58 MiMo-V2.5 1,175
  59. 59 Kimi K2.6 1,172
  60. 60 Claude Opus 4.5 1,171
  61. 61 DeepSeek V4 Flash 0731 1,171
  62. 62 Mercury 3 1,167
  63. 63 Kimi K2.7 Code 1,166
  64. 64 GLM-5 1,159
  65. 65 Grok 4.20 Beta reasoning 1,154
  66. 66 Qwen 3.6 Plus 1,152
  67. 67 MiniMax-M3 1,147
  68. 68 DeepSeek-V4-Flash 1,146
  69. 69 Claude Opus 4.1 thinking 1,143
  70. 70 Kimi K2.5 thinking 1,140
  71. 71 GLM-5V-Turbo 1,140
  72. 72 GPT-5.1 high 1,139
  73. 73 MiniMax-M2.5 1,139
  74. 74 GLM-4.7 1,133
  75. 75 Qwen3.5 397B-A17B 1,130
  76. 76 GPT-5.2 medium 1,129
  77. 77 DeepSeek-V4-Pro 1,129
  78. 78 MiniMax-M2.7 1,124
  79. 79 MiniMax-M2.1 1,122
  80. 80 GPT-5.3 Codex 1,119
  81. 81 Claude Opus 4 1,118
  82. 82 Claude Sonnet 4.5 thinking 1,105
  83. 83 Inkling 1,102
  84. 84 GLM-4.6 1,099
  85. 85 Qwen3.5 Plus 02-15 1,098
  86. 86 MiniMax M2 Stable 1,090
  87. 87 GLM-4.5 1,086
  88. 88 GPT-5 mini (Default) 1,081
  89. 89 Claude 3.7 Sonnet 1,078
  90. 90 Step 3.7 Flash 1,074
  91. 91 Grok 4.3 1,072
  92. 92 Claude Sonnet 4 thinking 1,069
  93. 93 GLM-4.5-Air 1,062
  94. 94 Nemotron 3 Ultra 1,061
  95. 95 DeepSeek-V3.1 thinking 1,055
  96. 96 v0-1.5-md 1,052
  97. 97 Gemini 3.1 Flash-Lite 1,043
  98. 98 DeepSeek-R1 May 2025 1,029
  99. 99 glm-4.7-flash 1,028
  100. 100 DeepSeek-V3.2 1,023
  101. 101 DeepSeek-V3.2-Exp 1,022
  102. 102 MiMo-V2-Flash 1,020
  103. 103 Claude Haiku 4.5 1,016
  104. 104 Grok 4 1,011
  105. 105 Gemini 2.5 Flash Sep 2025 1,010
  106. 106 Trinity Large Thinking 1,007
  107. 107 Qwen3-Max 1,002
  108. 108 Grok 4.1 Fast reasoning 999
  109. 109 Mercury 2 991
  110. 110 Mistral Medium 3.1 983
  111. 111 Mistral Large 3 983
  112. 112 Prime Intellect: INTELLECT-3 982
  113. 113 Grok 3 973
  114. 114 GPT-5.1-Codex-mini 968
  115. 115 DeepSeek-V3 Mar 2025 967
  116. 116 Grok 4 Fast reasoning 966
  117. 117 Mistral Medium 3.5 940
  118. 118 AesCoder-4B 921
  119. 119 Olmo 3.1 32B Think 862

Source: Official DesignArena SVG leaderboard

#EntryScoreSettingDateSource
1 Arrow 2 Telos 1,502 Official leaderboard
2 GPT-6 Astra 1,423 xhigh Sep 2026 Official leaderboard
3 Arrow 2 1,369 Official leaderboard
4 Gemini 4 Argon 1,335 Sep 2026 Official leaderboard
5 Kimi K3.1 Pro 1,327 Oct 2026 Official leaderboard
6 Claude Opus 5.5 1,321 Sep 2026 Official leaderboard
7 Muse Spark 1.3 1,316 xhigh, max Sep 2026 Official leaderboard
8 Claude Opus 5 1,307 Jul 2026 Official leaderboard
9 GPT-5.6 Sol 1,305 medium, xhigh Jul 2026 Official leaderboard
10 GPT-6 Sol 1,303 Sep 2026 Official leaderboard
11 Claude Sonnet 5.5 1,303 Sep 2026 Official leaderboard
12 Kimi K3 1,298 Jul 2026 Official leaderboard
13 Claude Fable 5 1,294 Jun 2026 Official leaderboard
14 Claude Fable 5.1 1,294 Sep 2026 Official leaderboard
15 GPT-6.1 Terra 1,294 Oct 2026 Official leaderboard
16 MiMo-V2.6 Pro 1,293 Sep 2026 Official leaderboard
17 Gemini 3.8 Flash 1,291 Sep 2026 Official leaderboard
18 MiniMax-M3.1-Flash 1,282 Oct 2026 Official leaderboard
19 Gemini 3.7 Flash 1,280 Aug 2026 Official leaderboard
20 Gemini 3.1 Pro 1,276 Feb 2026 Official leaderboard
21 Muse Spark 1.2 1,274 Aug 2026 Official leaderboard
22 GLM-5.3-Flash 1,272 Aug 2026 Official leaderboard
23 Arrow 1.1 Max 1,266 Official leaderboard
24 Ling-3.1-flash 1,266 Oct 2026 Official leaderboard
25 Arrow-1.0 1,265 Official leaderboard
26 GLM-5.3 1,248 Aug 2026 Official leaderboard
27 Gemini 3.5 Flash 1,245 May 2026 Official leaderboard
28 DeepSeek V4.1 Flash 1,244 1017 Sep 2026 Official leaderboard
29 GPT-6 Luna 1,238 Sep 2026 Official leaderboard
30 Gemini 3 Pro 1,235 Nov 2025 Official leaderboard
31 GPT-5.5 1,226 Apr 2026 Official leaderboard
32 Muse Spark 1.1 1,225 Jul 2026 Official leaderboard
33 Mistral Large 4 1,224 Oct 2026 Official leaderboard
34 Grok 4.6 1,221 Aug 2026 Official leaderboard
35 Claude Opus 4.6 1,218 thinking Feb 2026 Official leaderboard
36 Recraft V4.1 Vector Pro 1,217 Official leaderboard
37 Muse Spark 1,217 Apr 2026 Official leaderboard
38 Recraft V4 Vector 1,210 Official leaderboard
39 GLM-5.1 1,209 Apr 2026 Official leaderboard
40 GPT-5.6 Terra 1,209 Jul 2026 Official leaderboard
41 Recraft V4 Vector Pro 1,208 Official leaderboard
42 Qwen3.7-Max 1,207 May 2026 Official leaderboard
43 Qwen3.7-Plus 1,202 Jun 2026 Official leaderboard
44 Grok 4.5 1,201 Jul 2026 Official leaderboard
45 GLM-5.2 1,197 Jun 2026 Official leaderboard
46 Arrow 1.1 1,196 Official leaderboard
47 GLM-5-Turbo 1,195 Mar 2026 Official leaderboard
48 Tencent Hy3 1,194 Jul 2026 Official leaderboard
49 Claude Haiku 5.5 1,194 Oct 2026 Official leaderboard
50 Nex N2 Pro 1,187 Official leaderboard
51 Claude Sonnet 4.6 1,187 Feb 2026 Official leaderboard
52 GPT-5.6 Luna 1,187 Jul 2026 Official leaderboard
53 Claude Sonnet 5 1,186 Jun 2026 Official leaderboard
54 GPT-5.4 1,185 medium Mar 2026 Official leaderboard
55 MiMo-V2.5-Pro 1,183 Apr 2026 Official leaderboard
56 Claude Opus 4.8 1,183 May 2026 Official leaderboard
57 GPT-5 1,179 minimal, high Aug 2025 Official leaderboard
58 MiMo-V2.5 1,175 Apr 2026 Official leaderboard
59 Kimi K2.6 1,172 Apr 2026 Official leaderboard
60 Claude Opus 4.5 1,171 Nov 2025 Official leaderboard
61 DeepSeek V4 Flash 0731 1,171 Jul 2026 Official leaderboard
62 Mercury 3 1,167 Oct 2026 Official leaderboard
63 Kimi K2.7 Code 1,166 Jun 2026 Official leaderboard
64 GLM-5 1,159 Feb 2026 Official leaderboard
65 Grok 4.20 Beta 1,154 reasoning Official leaderboard
66 Qwen 3.6 Plus 1,152 Mar 2026 Official leaderboard
67 MiniMax-M3 1,147 Jun 2026 Official leaderboard
68 DeepSeek-V4-Flash 1,146 Apr 2026 Official leaderboard
69 Claude Opus 4.1 1,143 thinking Aug 2025 Official leaderboard
70 Kimi K2.5 1,140 thinking Jan 2026 Official leaderboard
71 GLM-5V-Turbo 1,140 Apr 2026 Official leaderboard
72 GPT-5.1 1,139 high Nov 2025 Official leaderboard
73 MiniMax-M2.5 1,139 Feb 2026 Official leaderboard
74 GLM-4.7 1,133 Dec 2025 Official leaderboard
75 Qwen3.5 397B-A17B 1,130 Feb 2026 Official leaderboard
76 GPT-5.2 1,129 medium Dec 2025 Official leaderboard
77 DeepSeek-V4-Pro 1,129 Apr 2026 Official leaderboard
78 MiniMax-M2.7 1,124 Mar 2026 Official leaderboard
79 MiniMax-M2.1 1,122 Dec 2025 Official leaderboard
80 GPT-5.3 Codex 1,119 Feb 2026 Official leaderboard
81 Claude Opus 4 1,118 May 2025 Official leaderboard
82 Claude Sonnet 4.5 1,105 thinking Sep 2025 Official leaderboard
83 Inkling 1,102 Jul 2026 Official leaderboard
84 GLM-4.6 1,099 Sep 2025 Official leaderboard
85 Qwen3.5 Plus 02-15 1,098 Official leaderboard
86 MiniMax M2 Stable 1,090 Official leaderboard
87 GLM-4.5 1,086 Aug 2025 Official leaderboard
88 GPT-5 mini (Default) 1,081 Official leaderboard
89 Claude 3.7 Sonnet 1,078 Feb 2025 Official leaderboard
90 Step 3.7 Flash 1,074 May 2026 Official leaderboard
91 Grok 4.3 1,072 Apr 2026 Official leaderboard
92 Claude Sonnet 4 1,069 thinking May 2025 Official leaderboard
93 GLM-4.5-Air 1,062 Jul 2025 Official leaderboard
94 Nemotron 3 Ultra 1,061 Jun 2026 Official leaderboard
95 DeepSeek-V3.1 1,055 thinking Aug 2025 Official leaderboard
96 v0-1.5-md 1,052 Official leaderboard
97 Gemini 3.1 Flash-Lite 1,043 Mar 2026 Official leaderboard
98 DeepSeek-R1 May 2025 1,029 May 2025 Official leaderboard
99 glm-4.7-flash 1,028 Jan 2026 Official leaderboard
100 DeepSeek-V3.2 1,023 Dec 2025 Official leaderboard
101 DeepSeek-V3.2-Exp 1,022 Sep 2025 Official leaderboard
102 MiMo-V2-Flash 1,020 Dec 2025 Official leaderboard
103 Claude Haiku 4.5 1,016 Oct 2025 Official leaderboard
104 Grok 4 1,011 Jul 2025 Official leaderboard
105 Gemini 2.5 Flash Sep 2025 1,010 Sep 2025 Official leaderboard
106 Trinity Large Thinking 1,007 Apr 2026 Official leaderboard
107 Qwen3-Max 1,002 Sep 2025 Official leaderboard
108 Grok 4.1 Fast 999 reasoning Nov 2025 Official leaderboard
109 Mercury 2 991 Feb 2026 Official leaderboard
110 Mistral Medium 3.1 983 Aug 2025 Official leaderboard
111 Mistral Large 3 983 Dec 2025 Official leaderboard
112 Prime Intellect: INTELLECT-3 982 Official leaderboard
113 Grok 3 973 Apr 2025 Official leaderboard
114 GPT-5.1-Codex-mini 968 Nov 2025 Official leaderboard
115 DeepSeek-V3 Mar 2025 967 Mar 2025 Official leaderboard
116 Grok 4 Fast 966 reasoning Sep 2025 Official leaderboard
117 Mistral Medium 3.5 940 Apr 2026 Official leaderboard
118 AesCoder-4B 921 Official leaderboard
119 Olmo 3.1 32B Think 862 Dec 2025 Official leaderboard

More effort adds a median 21 elo

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
GPT-6 Astra
xhigh, max
1,423
Muse Spark 1.3
xhigh
1,316
GPT-5.6 Sol
medium
1,305
DeepSeek V4.1 Flash
1017
1,244
Claude Opus 4.6
thinking
1,218
GPT-5.4
none medium
1,185
GPT-5
minimal minimal, hi…
1,179
Claude Opus 4.1
thinking
1,143
GPT-5.1
none high
1,139
GPT-5.2
none
1,129
Claude Sonnet 4.5
thinking
1,105
Claude Sonnet 4
thinking
1,069

Source: Official DesignArena SVG leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-6 Astra xhigh 1,423
GPT-6 Astra xhigh, max 1,416
Gemini 4 Argon 1,335
Kimi K3.1 Pro 1,327
Claude Opus 5.5 1,321
Muse Spark 1.3 xhigh, max 1,316
Claude Opus 5 1,307
GPT-5.6 Sol medium, xhigh 1,305
GPT-6 Sol 1,303
Claude Sonnet 5.5 1,303
Kimi K3 1,298
Muse Spark 1.3 xhigh 1,295
Claude Fable 5 1,294
Claude Fable 5.1 1,294
GPT-6.1 Terra 1,294
GPT-5.6 Sol medium 1,293
MiMo-V2.6 Pro 1,293
Gemini 3.8 Flash 1,291
MiniMax-M3.1-Flash 1,282
Gemini 3.7 Flash 1,280
Gemini 3.1 Pro 1,276
Muse Spark 1.2 1,274
GLM-5.3-Flash 1,272
Ling-3.1-flash 1,266
GLM-5.3 1,248
Gemini 3.5 Flash 1,245
DeepSeek V4.1 Flash 1017 1,244
GPT-6 Luna 1,238
Gemini 3 Pro 1,235
DeepSeek V4.1 Flash 1,232
GPT-5.5 1,226
Muse Spark 1.1 1,225
Mistral Large 4 1,224
Grok 4.6 1,221
Claude Opus 4.6 thinking 1,218
Muse Spark 1,217
GLM-5.1 1,209
GPT-5.6 Terra 1,209
Qwen3.7-Max 1,207
Claude Opus 4.6 1,205
Qwen3.7-Plus 1,202
Grok 4.5 1,201
GLM-5.2 1,197
GLM-5-Turbo 1,195
Tencent Hy3 1,194
Claude Haiku 5.5 1,194
Claude Sonnet 4.6 1,187
GPT-5.6 Luna 1,187
Claude Sonnet 5 1,186
GPT-5.4 medium 1,185
MiMo-V2.5-Pro 1,183
Claude Opus 4.8 1,183
GPT-5 minimal, high 1,179
MiMo-V2.5 1,175
Kimi K2.6 1,172
Claude Opus 4.5 1,171
DeepSeek V4 Flash 0731 1,171
Mercury 3 1,167
Kimi K2.7 Code 1,166
GLM-5 1,159
Qwen 3.6 Plus 1,152
MiniMax-M3 1,147
DeepSeek-V4-Flash 1,146
Claude Opus 4.1 thinking 1,143
Kimi K2.5 thinking 1,140
GLM-5V-Turbo 1,140
GPT-5.1 high 1,139
MiniMax-M2.5 1,139
GLM-4.7 1,133
GPT-5.4 low 1,132
Qwen3.5 397B-A17B 1,130
GPT-5.2 medium 1,129
DeepSeek-V4-Pro 1,129
GPT-5.2 low 1,126
MiniMax-M2.7 1,124
MiniMax-M2.1 1,122
GPT-5.1 medium 1,121
GPT-5.2 none 1,121
GPT-5.3 Codex 1,119
Claude Opus 4 1,118
GPT-5 minimal 1,118
GPT-5.4 none 1,117
GPT-5.1 low 1,108
Claude Sonnet 4.5 thinking 1,105
Inkling 1,102
Claude Opus 4.1 1,100
GLM-4.6 1,099
Claude Sonnet 4.5 1,098
GPT-5.1 none 1,094
GLM-4.5 1,086
Claude 3.7 Sonnet 1,078
Step 3.7 Flash 1,074
Grok 4.3 1,072
Claude Sonnet 4 thinking 1,069
GLM-4.5-Air 1,062
Nemotron 3 Ultra 1,061
DeepSeek-V3.1 thinking 1,055
Claude Sonnet 4 1,053
Gemini 3.1 Flash-Lite 1,043
DeepSeek-R1 May 2025 1,029
glm-4.7-flash 1,028
DeepSeek-V3.2 1,023
DeepSeek-V3.2-Exp 1,022
MiMo-V2-Flash 1,020
Claude Haiku 4.5 1,016
Grok 4 1,011
Gemini 2.5 Flash Sep 2025 1,010
Trinity Large Thinking 1,007
Qwen3-Max 1,002
Grok 4.1 Fast reasoning 999
Mercury 2 991
Mistral Medium 3.1 983
Mistral Large 3 983
Grok 3 973
GPT-5.1-Codex-mini 968
DeepSeek-V3 Mar 2025 967
Grok 4 Fast reasoning 966
DeepSeek-V3.1 957
Mistral Medium 3.5 940
Grok 4.1 Fast 934
Grok 4 Fast 929
Olmo 3.1 32B Think 862

11 models have held number one on the official leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

Anthropic OpenAI Google DeepMind Meta
Own leaderboard 11 leaders
GPT-6 Astra 1,423

Source: Official DesignArena SVG leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard
800 1,000 1,200 1,400 1,600 2025 2026 2027 800 1,000 1,200 1,400 1,600 2025 2026 2027

Source: Official DesignArena SVG leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Coding

All 282 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471