Benchmark Almanac

DesignArena 3D Design

GPT-6 Astra 1,525 leads since September 2026
In use

Models build an interactive 3D scene that runs in the browser as a single HTML file. Users explore the scenes and vote for the better one.

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
Claude Fable 5.2 #1 on the capability index
Last in a launch post
November 2026
Last model added
November 2026
Scores refreshed
November 12th, 2026
156
Entries on its own leaderboard, 136 models we track
1
Launch post lists it, from 1 lab

GPT-6 Astra leads with 1,525

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.

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

  1. 1,525 GPT-6 Astra
  2. 1,516 Claude Fable 5.2
  3. 1,464 Claude Opus 5.5
  4. 1,460 Inkling 2 Small
  5. 1,446 Kimi K3.1 Pro
  6. 1,434 Muse Wonder
  7. 1,428 GPT-6.1 Sol
  8. 1,419 Gemini 4 Carbon
  9. 1,401 Claude Sonnet 5.5
  10. 1,391 Claude Fable 5.1
  11. 1,383 Muse Spark 1.3
  12. 1,374 Muse Ember
  13. 1,372 GPT-5.6 Sol
  14. 1,367 Kimi K3
  15. 1,356 Gemini 4 Argon
  16. 1,347 DeepSeek V4.1 Flash
  17. 1,327 Claude Haiku 5.5
  18. 1,324 Claude Opus 5
  19. 1,324 Ling-3.1-flash
  20. 1,323 MiMo-V2.6 Pro
  21. 1,321 Qwen 3.8 Max
  22. 1,318 MiniMax-M3.1-Flash
  23. 1,305 GPT-6.1 Terra
  24. 1,300 GPT-6 Sol
  25. 1,299 Claude Fable 5
  26. 1,299 GLM-5.3
  27. 1,296 GLM-5.3-Flash
  28. 1,291 Gemini 3.7 Flash
  29. 1,286 Muse Spark 1.2
  30. 1,278 Gemini 3.8 Flash
  31. 1,269 GLM-5.2
  32. 1,261 Deca 2.6 Ultra
  33. 1,259 Grok 4.7
  34. 1,258 Tencent Hy4
  35. 1,257 Claude Opus 4.6
  36. 1,257 Grok 4.6
  37. 1,257 Muse Spark Open
  38. 1,256 Deca 2.6 Pro
  39. 1,253 Gemini 3.6 Flash
  40. 1,250 Kimi K2.6
  41. 1,248 Qwen3.7-Max
  42. 1,242 Mistral Large 4
  43. 1,241 Muse Spark 1.1
  44. 1,240 Claude Sonnet 5
  45. 1,238 Grok 4.5
  46. 1,235 MiMo-V2.5-Pro
  47. 1,235 DeepSeek-V4-Pro
  48. 1,234 GLM-5.1
  49. 1,231 Nex N2 Pro
  50. 1,231 GLM-5-Turbo
  51. 1,227 Gemini 3.5 Flash
  52. 1,224 Mercury 3
  53. 1,223 Kimi K2.7 Code
  54. 1,222 Deca 2.5 Ultra
  55. 1,221 Qwen3.7-Plus
  56. 1,217 Claude Sonnet 4.6
  57. 1,217 Gemini 3.1 Pro
  58. 1,212 GLM-5
  59. 1,212 Tencent Hy3
  60. 1,206 Gemini 3 Pro
  61. 1,205 Claude Opus 4.8
  62. 1,203 Muse Spark
  63. 1,200 MiMo-V2.5
  64. 1,198 GPT-6 Luna
  65. 1,196 Claude Opus 4.5
  66. 1,192 GLM-5V-Turbo
  67. 1,192 MiniMax-M3
  68. 1,190 DeepSeek V4 Flash 0731
  69. 1,188 Kimi K2.5
  70. 1,186 Qwen 3.6 Plus
  71. 1,178 MiniMax-M2.7
  72. 1,178 GPT-5.5
  73. 1,176 Grok 4.20 Beta
  74. 1,173 GLM-4.7
  75. 1,171 DeepSeek-V4-Flash
  76. 1,166 Gemini 3 Flash
  77. 1,162 Solar Pro 4
  78. 1,153 GLM-4.5
  79. 1,153 MiniMax-M2.5
  80. 1,148 MiniMax-M2.1
  81. 1,145 Inkling
  82. 1,144 Qwen3.5 397B-A17B
  83. 1,140 Claude Sonnet 4.5
  84. 1,136 Claude Opus 4.1
  85. 1,131 DeepSeek-V3.2-Exp
  86. 1,123 DeepSeek-V3.1
  87. 1,123 Beam 1.0
  88. 1,122 AGI-01 Swift
  89. 1,122 Claude Sonnet 4
  90. 1,121 Claude Opus 4
  91. 1,117 Nemotron 3.5 Super VL
  92. 1,115 DeepSeek-V3.2
  93. 1,112 GPT-5.6 Terra
  94. 1,111 GLM-4.6
  95. 1,111 Step 3.7 Flash
  96. 1,111 Nemotron 3 Ultra
  97. 1,107 Grok 4.3
  98. 1,105 GLM-4.5-Air
  99. 1,103 glm-4.7-flash
  100. 1,102 Qwen3.5 Plus 02-15
  101. 1,102 GPT-5.6 Luna
  102. 1,098 MiMo-V2-Flash
  103. 1,094 DeepSeek-R1
  104. 1,092 Solar Mini 4
  105. 1,084 GPT-5.4
  106. 1,082 Mistral Large 3
  107. 1,080 MiniMax M2 Stable
  108. 1,077 Claude 3.7 Sonnet
  109. 1,073 v0-1.5-md
  110. 1,070 Muse Glimmer 30B
  111. 1,069 Prime Intellect: INTELLECT-3
  112. 1,069 DeepSeek-V3
  113. 1,066 Mistral Medium 3 (2505)
  114. 1,064 AesCoder-4B
  115. 1,064 Mistral Medium 3.1
  116. 1,064 Trinity Large Thinking
  117. 1,063 Gemini 2.5 Pro
  118. 1,062 GPT-5.2
  119. 1,057 Qwen3-Max
  120. 1,055 Claude Haiku 4.5
  121. 1,054 Gemini 2.5 Flash
  122. 1,048 grok-code-fast-1
  123. 1,044 GPT-5.1
  124. 1,044 Grok 4.1 Fast
  125. 1,043 Mistral Medium 3.5
  126. 1,039 Grok 4 Fast
  127. 1,038 GPT-5
  128. 1,034 Grok 4
  129. 1,032 Grok 3
  130. 1,031 Gemini 3.1 Flash-Lite
  131. 1,020 GPT-5 mini (Default)
  132. 1,016 Magistral Medium 1.2 (2509)
  133. 1,015 Devstral Medium
  134. 1,014 Ministral 3 8B Instruct
  135. 1,003 Codestral 2508
  136. 991 GPT-5.3 Codex
  137. 990 Magistral Small 1.2 (2509)
  138. 985 Olmo 3.1 32B Think
  139. 980 Qwen3-235B-A22B-Thinking
  140. 977 Qwen3-235B-A22B-Instruct-2507
  141. 973 Ministral 3 14B
  142. 971 GPT-5.1-Codex-mini
  143. 965 Llama 3.1 Nemotron Ultra 253B
  144. 956 Mercury 2
  145. 948 GPT-5 nano (Default)
  146. 948 Gemini 2.5 Flash-Lite
  147. 945 Ministral 3 3B Instruct
  148. 938 Grok-3 mini
  149. 908 GPT-4.1 nano
  150. 884 Llama 4 Maverick
  151. 884 gpt-oss-120b
  152. 852 GPT-4o
  153. 841 Qwen3-235B-A22B
  154. 837 o4-mini
  155. 833 GPT-4.1
  156. 819 GPT-4.1 mini

Bars start at 800.

  1. 1 GPT-6 Astra xhigh, max 1,525
  2. 2 Claude Fable 5.2 1,516
  3. 3 Claude Opus 5.5 1,464
  4. 4 Inkling 2 Small 1,460
  5. 5 Kimi K3.1 Pro 1,446
  6. 6 Muse Wonder max 1,434
  7. 7 GPT-6.1 Sol 1,428
  8. 8 Gemini 4 Carbon 1,419
  9. 9 Claude Sonnet 5.5 1,401
  10. 10 Claude Fable 5.1 1,391
  11. 11 Muse Spark 1.3 xhigh, max 1,383
  12. 12 Muse Ember 1,374
  13. 13 GPT-5.6 Sol medium, xhigh 1,372
  14. 14 Kimi K3 1,367
  15. 15 Gemini 4 Argon 1,356
  16. 16 DeepSeek V4.1 Flash 1017 1,347
  17. 17 Claude Haiku 5.5 1,327
  18. 18 Claude Opus 5 1,324
  19. 19 Ling-3.1-flash 1,324
  20. 20 MiMo-V2.6 Pro 1,323
  21. 21 Qwen 3.8 Max 1,321
  22. 22 MiniMax-M3.1-Flash 1,318
  23. 23 GPT-6.1 Terra 1,305
  24. 24 GPT-6 Sol 1,300
  25. 25 Claude Fable 5 1,299
  26. 26 GLM-5.3 1,299
  27. 27 GLM-5.3-Flash 1,296
  28. 28 Gemini 3.7 Flash 1,291
  29. 29 Muse Spark 1.2 1,286
  30. 30 Gemini 3.8 Flash 1,278
  31. 31 GLM-5.2 1,269
  32. 32 Deca 2.6 Ultra 1,261
  33. 33 Grok 4.7 1,259
  34. 34 Tencent Hy4 1,258
  35. 35 Claude Opus 4.6 1,257
  36. 36 Grok 4.6 1,257
  37. 37 Muse Spark Open xhigh 1,257
  38. 38 Deca 2.6 Pro 1,256
  39. 39 Gemini 3.6 Flash 1,253
  40. 40 Kimi K2.6 1,250
  41. 41 Qwen3.7-Max 1,248
  42. 42 Mistral Large 4 1,242
  43. 43 Muse Spark 1.1 1,241
  44. 44 Claude Sonnet 5 1,240
  45. 45 Grok 4.5 1,238
  46. 46 MiMo-V2.5-Pro 1,235
  47. 47 DeepSeek-V4-Pro 1,235
  48. 48 GLM-5.1 1,234
  49. 49 Nex N2 Pro 1,231
  50. 50 GLM-5-Turbo 1,231
  51. 51 Gemini 3.5 Flash 1,227
  52. 52 Mercury 3 1,224
  53. 53 Kimi K2.7 Code 1,223
  54. 54 Deca 2.5 Ultra 1,222
  55. 55 Qwen3.7-Plus 1,221
  56. 56 Claude Sonnet 4.6 1,217
  57. 57 Gemini 3.1 Pro 1,217
  58. 58 GLM-5 1,212
  59. 59 Tencent Hy3 1,212
  60. 60 Gemini 3 Pro 1,206
  61. 61 Claude Opus 4.8 1,205
  62. 62 Muse Spark 1,203
  63. 63 MiMo-V2.5 1,200
  64. 64 GPT-6 Luna 1,198
  65. 65 Claude Opus 4.5 1,196
  66. 66 GLM-5V-Turbo 1,192
  67. 67 MiniMax-M3 1,192
  68. 68 DeepSeek V4 Flash 0731 1,190
  69. 69 Kimi K2.5 thinking 1,188
  70. 70 Qwen 3.6 Plus 1,186
  71. 71 MiniMax-M2.7 1,178
  72. 72 GPT-5.5 1,178
  73. 73 Grok 4.20 Beta reasoning 1,176
  74. 74 GLM-4.7 1,173
  75. 75 DeepSeek-V4-Flash 1,171
  76. 76 Gemini 3 Flash 1,166
  77. 77 Solar Pro 4 1,162
  78. 78 GLM-4.5 1,153
  79. 79 MiniMax-M2.5 1,153
  80. 80 MiniMax-M2.1 1,148
  81. 81 Inkling 1,145
  82. 82 Qwen3.5 397B-A17B 1,144
  83. 83 Claude Sonnet 4.5 1,140
  84. 84 Claude Opus 4.1 1,136
  85. 85 DeepSeek-V3.2-Exp 1,131
  86. 86 DeepSeek-V3.1 thinking 1,123
  87. 87 Beam 1.0 max 1,123
  88. 88 AGI-01 Swift 1,122
  89. 89 Claude Sonnet 4 1,122
  90. 90 Claude Opus 4 1,121
  91. 91 Nemotron 3.5 Super VL 1,117
  92. 92 DeepSeek-V3.2 1,115
  93. 93 GPT-5.6 Terra 1,112
  94. 94 GLM-4.6 1,111
  95. 95 Step 3.7 Flash 1,111
  96. 96 Nemotron 3 Ultra 1,111
  97. 97 Grok 4.3 1,107
  98. 98 GLM-4.5-Air 1,105
  99. 99 glm-4.7-flash 1,103
  100. 100 Qwen3.5 Plus 02-15 1,102
  101. 101 GPT-5.6 Luna 1,102
  102. 102 MiMo-V2-Flash 1,098
  103. 103 DeepSeek-R1 May 2025 1,094
  104. 104 Solar Mini 4 1,092
  105. 105 GPT-5.4 medium 1,084
  106. 106 Mistral Large 3 1,082
  107. 107 MiniMax M2 Stable 1,080
  108. 108 Claude 3.7 Sonnet 1,077
  109. 109 v0-1.5-md 1,073
  110. 110 Muse Glimmer 30B xhigh 1,070
  111. 111 Prime Intellect: INTELLECT-3 1,069
  112. 112 DeepSeek-V3 Mar 2025 1,069
  113. 113 Mistral Medium 3 (2505) 1,066
  114. 114 AesCoder-4B 1,064
  115. 115 Mistral Medium 3.1 1,064
  116. 116 Trinity Large Thinking 1,064
  117. 117 Gemini 2.5 Pro Jun 2025 1,063
  118. 118 GPT-5.2 medium 1,062
  119. 119 Qwen3-Max 1,057
  120. 120 Claude Haiku 4.5 1,055
  121. 121 Gemini 2.5 Flash Sep 2025 1,054
  122. 122 grok-code-fast-1 1,048
  123. 123 GPT-5.1 high 1,044
  124. 124 Grok 4.1 Fast 1,044
  125. 125 Mistral Medium 3.5 1,043
  126. 126 Grok 4 Fast 1,039
  127. 127 GPT-5 minimal 1,038
  128. 128 Grok 4 1,034
  129. 129 Grok 3 1,032
  130. 130 Gemini 3.1 Flash-Lite 1,031
  131. 131 GPT-5 mini (Default) 1,020
  132. 132 Magistral Medium 1.2 (2509) 1,016
  133. 133 Devstral Medium 1,015
  134. 134 Ministral 3 8B Instruct 1,014
  135. 135 Codestral 2508 1,003
  136. 136 GPT-5.3 Codex 991
  137. 137 Magistral Small 1.2 (2509) 990
  138. 138 Olmo 3.1 32B Think 985
  139. 139 Qwen3-235B-A22B-Thinking Jul 2025 980
  140. 140 Qwen3-235B-A22B-Instruct-2507 977
  141. 141 Ministral 3 14B 973
  142. 142 GPT-5.1-Codex-mini 971
  143. 143 Llama 3.1 Nemotron Ultra 253B 965
  144. 144 Mercury 2 956
  145. 145 GPT-5 nano (Default) 948
  146. 146 Gemini 2.5 Flash-Lite Sep 2025 948
  147. 147 Ministral 3 3B Instruct 945
  148. 148 Grok-3 mini 938
  149. 149 GPT-4.1 nano 908
  150. 150 Llama 4 Maverick 884
  151. 151 gpt-oss-120b 884
  152. 152 GPT-4o Mar 2025 852
  153. 153 Qwen3-235B-A22B 841
  154. 154 o4-mini 837
  155. 155 GPT-4.1 833
  156. 156 GPT-4.1 mini 819

Source: Official DesignArena 3D Design leaderboard

#EntryScoreSettingDateSource
1 GPT-6 Astra 1,525 xhigh, max Sep 2026 Official leaderboard
2 Claude Fable 5.2 1,516 Nov 2026 Official leaderboard
3 Claude Opus 5.5 1,464 Sep 2026 Official leaderboard
4 Inkling 2 Small 1,460 Nov 2026 Official leaderboard
5 Kimi K3.1 Pro 1,446 Oct 2026 Official leaderboard
6 Muse Wonder 1,434 max Oct 2026 Official leaderboard
7 GPT-6.1 Sol 1,428 Sep 2026 Official leaderboard
8 Gemini 4 Carbon 1,419 Oct 2026 Official leaderboard
9 Claude Sonnet 5.5 1,401 Sep 2026 Official leaderboard
10 Claude Fable 5.1 1,391 Sep 2026 Official leaderboard
11 Muse Spark 1.3 1,383 xhigh, max Sep 2026 Official leaderboard
12 Muse Ember 1,374 Nov 2026 Official leaderboard
13 GPT-5.6 Sol 1,372 medium, xhigh Jul 2026 Official leaderboard
14 Kimi K3 1,367 Jul 2026 Official leaderboard
15 Gemini 4 Argon 1,356 Sep 2026 Official leaderboard
16 DeepSeek V4.1 Flash 1,347 1017 Sep 2026 Official leaderboard
17 Claude Haiku 5.5 1,327 Oct 2026 Official leaderboard
18 Claude Opus 5 1,324 Jul 2026 Official leaderboard
19 Ling-3.1-flash 1,324 Oct 2026 Official leaderboard
20 MiMo-V2.6 Pro 1,323 Sep 2026 Official leaderboard
21 Qwen 3.8 Max 1,321 Aug 2026 Official leaderboard
22 MiniMax-M3.1-Flash 1,318 Oct 2026 Official leaderboard
23 GPT-6.1 Terra 1,305 Oct 2026 Official leaderboard
24 GPT-6 Sol 1,300 Sep 2026 Official leaderboard
25 Claude Fable 5 1,299 Jun 2026 Official leaderboard
26 GLM-5.3 1,299 Aug 2026 Official leaderboard
27 GLM-5.3-Flash 1,296 Aug 2026 Official leaderboard
28 Gemini 3.7 Flash 1,291 Aug 2026 Official leaderboard
29 Muse Spark 1.2 1,286 Aug 2026 Official leaderboard
30 Gemini 3.8 Flash 1,278 Sep 2026 Official leaderboard
31 GLM-5.2 1,269 Jun 2026 Official leaderboard
32 Deca 2.6 Ultra 1,261 Official leaderboard
33 Grok 4.7 1,259 Sep 2026 Official leaderboard
34 Tencent Hy4 1,258 Nov 2026 Official leaderboard
35 Claude Opus 4.6 1,257 Feb 2026 Official leaderboard
36 Grok 4.6 1,257 Aug 2026 Official leaderboard
37 Muse Spark Open 1,257 xhigh Oct 2026 Official leaderboard
38 Deca 2.6 Pro 1,256 Official leaderboard
39 Gemini 3.6 Flash 1,253 Jul 2026 Official leaderboard
40 Kimi K2.6 1,250 Apr 2026 Official leaderboard
41 Qwen3.7-Max 1,248 May 2026 Official leaderboard
42 Mistral Large 4 1,242 Oct 2026 Official leaderboard
43 Muse Spark 1.1 1,241 Jul 2026 Official leaderboard
44 Claude Sonnet 5 1,240 Jun 2026 Official leaderboard
45 Grok 4.5 1,238 Jul 2026 Official leaderboard
46 MiMo-V2.5-Pro 1,235 Apr 2026 Official leaderboard
47 DeepSeek-V4-Pro 1,235 Apr 2026 Official leaderboard
48 GLM-5.1 1,234 Apr 2026 Official leaderboard
49 Nex N2 Pro 1,231 Official leaderboard
50 GLM-5-Turbo 1,231 Mar 2026 Official leaderboard
51 Gemini 3.5 Flash 1,227 May 2026 Official leaderboard
52 Mercury 3 1,224 Oct 2026 Official leaderboard
53 Kimi K2.7 Code 1,223 Jun 2026 Official leaderboard
54 Deca 2.5 Ultra 1,222 Official leaderboard
55 Qwen3.7-Plus 1,221 Jun 2026 Official leaderboard
56 Claude Sonnet 4.6 1,217 Feb 2026 Official leaderboard
57 Gemini 3.1 Pro 1,217 Feb 2026 Official leaderboard
58 GLM-5 1,212 Feb 2026 Official leaderboard
59 Tencent Hy3 1,212 Jul 2026 Official leaderboard
60 Gemini 3 Pro 1,206 Nov 2025 Official leaderboard
61 Claude Opus 4.8 1,205 May 2026 Official leaderboard
62 Muse Spark 1,203 Apr 2026 Official leaderboard
63 MiMo-V2.5 1,200 Apr 2026 Official leaderboard
64 GPT-6 Luna 1,198 Sep 2026 Official leaderboard
65 Claude Opus 4.5 1,196 Nov 2025 Official leaderboard
66 GLM-5V-Turbo 1,192 Apr 2026 Official leaderboard
67 MiniMax-M3 1,192 Jun 2026 Official leaderboard
68 DeepSeek V4 Flash 0731 1,190 Jul 2026 Official leaderboard
69 Kimi K2.5 1,188 thinking Jan 2026 Official leaderboard
70 Qwen 3.6 Plus 1,186 Mar 2026 Official leaderboard
71 MiniMax-M2.7 1,178 Mar 2026 Official leaderboard
72 GPT-5.5 1,178 Apr 2026 Official leaderboard
73 Grok 4.20 Beta 1,176 reasoning Official leaderboard
74 GLM-4.7 1,173 Dec 2025 Official leaderboard
75 DeepSeek-V4-Flash 1,171 Apr 2026 Official leaderboard
76 Gemini 3 Flash 1,166 Dec 2025 Official leaderboard
77 Solar Pro 4 1,162 Aug 2026 Official leaderboard
78 GLM-4.5 1,153 Aug 2025 Official leaderboard
79 MiniMax-M2.5 1,153 Feb 2026 Official leaderboard
80 MiniMax-M2.1 1,148 Dec 2025 Official leaderboard
81 Inkling 1,145 Jul 2026 Official leaderboard
82 Qwen3.5 397B-A17B 1,144 Feb 2026 Official leaderboard
83 Claude Sonnet 4.5 1,140 Sep 2025 Official leaderboard
84 Claude Opus 4.1 1,136 Aug 2025 Official leaderboard
85 DeepSeek-V3.2-Exp 1,131 Sep 2025 Official leaderboard
86 DeepSeek-V3.1 1,123 thinking Aug 2025 Official leaderboard
87 Beam 1.0 1,123 max Oct 2026 Official leaderboard
88 AGI-01 Swift 1,122 Official leaderboard
89 Claude Sonnet 4 1,122 May 2025 Official leaderboard
90 Claude Opus 4 1,121 May 2025 Official leaderboard
91 Nemotron 3.5 Super VL 1,117 Oct 2026 Official leaderboard
92 DeepSeek-V3.2 1,115 Dec 2025 Official leaderboard
93 GPT-5.6 Terra 1,112 Jul 2026 Official leaderboard
94 GLM-4.6 1,111 Sep 2025 Official leaderboard
95 Step 3.7 Flash 1,111 May 2026 Official leaderboard
96 Nemotron 3 Ultra 1,111 Jun 2026 Official leaderboard
97 Grok 4.3 1,107 Apr 2026 Official leaderboard
98 GLM-4.5-Air 1,105 Jul 2025 Official leaderboard
99 glm-4.7-flash 1,103 Jan 2026 Official leaderboard
100 Qwen3.5 Plus 02-15 1,102 Official leaderboard
101 GPT-5.6 Luna 1,102 Jul 2026 Official leaderboard
102 MiMo-V2-Flash 1,098 Dec 2025 Official leaderboard
103 DeepSeek-R1 May 2025 1,094 May 2025 Official leaderboard
104 Solar Mini 4 1,092 Oct 2026 Official leaderboard
105 GPT-5.4 1,084 medium Mar 2026 Official leaderboard
106 Mistral Large 3 1,082 Dec 2025 Official leaderboard
107 MiniMax M2 Stable 1,080 Official leaderboard
108 Claude 3.7 Sonnet 1,077 Feb 2025 Official leaderboard
109 v0-1.5-md 1,073 Official leaderboard
110 Muse Glimmer 30B 1,070 xhigh Official leaderboard
111 Prime Intellect: INTELLECT-3 1,069 Official leaderboard
112 DeepSeek-V3 Mar 2025 1,069 Mar 2025 Official leaderboard
113 Mistral Medium 3 (2505) 1,066 Official leaderboard
114 AesCoder-4B 1,064 Official leaderboard
115 Mistral Medium 3.1 1,064 Aug 2025 Official leaderboard
116 Trinity Large Thinking 1,064 Apr 2026 Official leaderboard
117 Gemini 2.5 Pro Jun 2025 1,063 Jun 2025 Official leaderboard
118 GPT-5.2 1,062 medium Dec 2025 Official leaderboard
119 Qwen3-Max 1,057 Sep 2025 Official leaderboard
120 Claude Haiku 4.5 1,055 Oct 2025 Official leaderboard
121 Gemini 2.5 Flash Sep 2025 1,054 Sep 2025 Official leaderboard
122 grok-code-fast-1 1,048 Aug 2025 Official leaderboard
123 GPT-5.1 1,044 high Nov 2025 Official leaderboard
124 Grok 4.1 Fast 1,044 Nov 2025 Official leaderboard
125 Mistral Medium 3.5 1,043 Apr 2026 Official leaderboard
126 Grok 4 Fast 1,039 Sep 2025 Official leaderboard
127 GPT-5 1,038 minimal Aug 2025 Official leaderboard
128 Grok 4 1,034 Jul 2025 Official leaderboard
129 Grok 3 1,032 Apr 2025 Official leaderboard
130 Gemini 3.1 Flash-Lite 1,031 Mar 2026 Official leaderboard
131 GPT-5 mini (Default) 1,020 Official leaderboard
132 Magistral Medium 1.2 (2509) 1,016 Official leaderboard
133 Devstral Medium 1,015 Jul 2025 Official leaderboard
134 Ministral 3 8B Instruct 1,014 Dec 2025 Official leaderboard
135 Codestral 2508 1,003 Official leaderboard
136 GPT-5.3 Codex 991 Feb 2026 Official leaderboard
137 Magistral Small 1.2 (2509) 990 Official leaderboard
138 Olmo 3.1 32B Think 985 Dec 2025 Official leaderboard
139 Qwen3-235B-A22B-Thinking Jul 2025 980 Jul 2025 Official leaderboard
140 Qwen3-235B-A22B-Instruct-2507 977 Official leaderboard
141 Ministral 3 14B 973 Dec 2025 Official leaderboard
142 GPT-5.1-Codex-mini 971 Nov 2025 Official leaderboard
143 Llama 3.1 Nemotron Ultra 253B 965 Official leaderboard
144 Mercury 2 956 Feb 2026 Official leaderboard
145 GPT-5 nano (Default) 948 Official leaderboard
146 Gemini 2.5 Flash-Lite Sep 2025 948 Sep 2025 Official leaderboard
147 Ministral 3 3B Instruct 945 Dec 2025 Official leaderboard
148 Grok-3 mini 938 Apr 2025 Official leaderboard
149 GPT-4.1 nano 908 Apr 2025 Official leaderboard
150 Llama 4 Maverick 884 Apr 2025 Official leaderboard
151 gpt-oss-120b 884 Aug 2025 Official leaderboard
152 GPT-4o Mar 2025 852 Mar 2025 Official leaderboard
153 Qwen3-235B-A22B 841 Apr 2025 Official leaderboard
154 o4-mini 837 Apr 2025 Official leaderboard
155 GPT-4.1 833 Apr 2025 Official leaderboard
156 GPT-4.1 mini 819 Apr 2025 Official leaderboard

Labs report scores for 1 model in their cards. We plot every source in the charts below.

More effort adds a median 33 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 xhigh, max
1,525
Muse Spark 1.3
xhigh
1,383
GPT-5.6 Sol
medium medium, xhi…
1,372
DeepSeek V4.1 Flash
1017
1,347
Claude Opus 4.6
thinking
1,257
Tencent Hy3
low
1,212
Claude Sonnet 4.5
thinking
1,140
Claude Opus 4.1
thinking
1,136
DeepSeek-V3.1
thinking
1,123
Claude Sonnet 4
thinking
1,122
GPT-5.4
none medium
1,084
GPT-5.2
none
1,062

Source: Official DesignArena 3D Design leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-6 Astra xhigh, max 1,525
Claude Fable 5.2 1,516
Claude Opus 5.5 1,464
GPT-6 Astra xhigh 1,462
Inkling 2 Small 1,460
Kimi K3.1 Pro 1,446
Muse Wonder max 1,434
GPT-6.1 Sol 1,428
Gemini 4 Carbon 1,419
Claude Sonnet 5.5 1,401
Claude Fable 5.1 1,391
Muse Spark 1.3 xhigh, max 1,383
Muse Ember 1,374
GPT-5.6 Sol medium, xhigh 1,372
Kimi K3 1,367
Gemini 4 Argon 1,356
Muse Spark 1.3 xhigh 1,350
DeepSeek V4.1 Flash 1017 1,347
DeepSeek V4.1 Flash 1,342
Claude Haiku 5.5 1,327
Claude Opus 5 1,324
Ling-3.1-flash 1,324
MiMo-V2.6 Pro 1,323
Qwen 3.8 Max 1,321
MiniMax-M3.1-Flash 1,318
GPT-6.1 Terra 1,305
GPT-6 Sol 1,300
Claude Fable 5 1,299
GLM-5.3 1,299
GLM-5.3-Flash 1,296
GPT-5.6 Sol medium 1,295
Gemini 3.7 Flash 1,291
Muse Spark 1.2 1,286
Gemini 3.8 Flash 1,278
GLM-5.2 1,269
Grok 4.7 1,259
Tencent Hy4 1,258
Claude Opus 4.6 1,257
Grok 4.6 1,257
Muse Spark Open xhigh 1,257
Claude Opus 4.6 thinking 1,254
Gemini 3.6 Flash 1,253
Kimi K2.6 1,250
Qwen3.7-Max 1,248
Mistral Large 4 1,242
Muse Spark 1.1 1,241
Claude Sonnet 5 1,240
Grok 4.5 1,238
MiMo-V2.5-Pro 1,235
DeepSeek-V4-Pro 1,235
GLM-5.1 1,234
GLM-5-Turbo 1,231
Gemini 3.5 Flash 1,227
Mercury 3 1,224
Kimi K2.7 Code 1,223
Qwen3.7-Plus 1,221
Claude Sonnet 4.6 1,217
Gemini 3.1 Pro 1,217
GLM-5 1,212
Tencent Hy3 1,212
Gemini 3 Pro 1,206
Claude Opus 4.8 1,205
Muse Spark 1,203
MiMo-V2.5 1,200
GPT-6 Luna 1,198
Claude Opus 4.5 1,196
GLM-5V-Turbo 1,192
MiniMax-M3 1,192
DeepSeek V4 Flash 0731 1,190
Kimi K2.5 thinking 1,188
Qwen 3.6 Plus 1,186
MiniMax-M2.7 1,178
GPT-5.5 1,178
GLM-4.7 1,173
DeepSeek-V4-Flash 1,171
Gemini 3 Flash 1,166
Solar Pro 4 1,162
Tencent Hy3 low 1,157
GLM-4.5 1,153
MiniMax-M2.5 1,153
MiniMax-M2.1 1,148
Inkling 1,145
Qwen3.5 397B-A17B 1,144
Claude Sonnet 4.5 1,140
Claude Opus 4.1 1,136
Claude Sonnet 4.5 thinking 1,135
DeepSeek-V3.2-Exp 1,131
Claude Opus 4.1 thinking 1,123
DeepSeek-V3.1 thinking 1,123
Beam 1.0 max 1,123
Claude Sonnet 4 1,122
Claude Opus 4 1,121
Nemotron 3.5 Super VL 1,117
DeepSeek-V3.2 1,115
GPT-5.6 Terra 1,112
GLM-4.6 1,111
Step 3.7 Flash 1,111
Nemotron 3 Ultra 1,111
Claude Sonnet 4 thinking 1,109
Grok 4.3 1,107
GLM-4.5-Air 1,105
glm-4.7-flash 1,103
GPT-5.6 Luna 1,102
MiMo-V2-Flash 1,098
DeepSeek-R1 May 2025 1,094
Solar Mini 4 1,092
GPT-5.4 medium 1,084
Mistral Large 3 1,082
Claude 3.7 Sonnet 1,077
DeepSeek-V3 Mar 2025 1,069
Mistral Medium 3.1 1,064
Trinity Large Thinking 1,064
Gemini 2.5 Pro Jun 2025 1,063
GPT-5.2 medium 1,062
DeepSeek-V3.1 1,060
Qwen3-Max 1,057
GPT-5.2 none 1,056
Claude Haiku 4.5 1,055
Gemini 2.5 Flash Sep 2025 1,054
GPT-5.2 low 1,049
grok-code-fast-1 1,048
GPT-5.4 low 1,048
GPT-5.1 high 1,044
Grok 4.1 Fast 1,044
Mistral Medium 3.5 1,043
Grok 4 Fast 1,039
GPT-5 minimal 1,038
GPT-5.2 high 1,036
Grok 4 1,034
Grok 4 Fast reasoning 1,033
GPT-5.1 none 1,033
Grok 3 1,032
Gemini 3.1 Flash-Lite 1,031
GPT-5.4 none 1,030
GPT-5.1 low 1,025
GPT-5.1 medium 1,021
Devstral Medium 1,015
Ministral 3 8B Instruct 1,014
GPT-5 minimal, high 1,009
Grok 4.1 Fast reasoning 1,005
Gemini 2.5 Flash Sep 2025 1,004
GPT-5.3 Codex 991
Olmo 3.1 32B Think 985
Qwen3-235B-A22B-Thinking Jul 2025 980
Ministral 3 14B 973
GPT-5.1-Codex-mini 971
Mercury 2 956
Gemini 2.5 Flash-Lite Sep 2025 948
Ministral 3 3B Instruct 945
Grok-3 mini 938
GPT-4.1 nano 908
Llama 4 Maverick 884
gpt-oss-120b 884
GPT-4o Mar 2025 852
Qwen3-235B-A22B 841
o4-mini 837
GPT-4.1 833
GPT-4.1 mini 819

9 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 Z.ai Google DeepMind OpenAI
Own leaderboard 9 leaders
GPT-6 Astra 1,525

Source: Official DesignArena 3D Design 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 Reported by the lab in its model card
800 1,000 1,200 1,400 1,600 2025 2026 2027 800 1,000 1,200 1,400 1,600 2025 2026 2027

Sources: Official DesignArena 3D Design leaderboard and labs’ model cards

One lab put it in a launch post in the last six months

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

Thinking Machines
DateLaunch postScoreAs the post writes it
Nov 5, 2026 Inkling 2 Small 1401 DesignArena 3D Design

One lab claim next to the leaderboard

Each score a lab reported for its own model, beside the same model's best score on the leaderboard. Gaps mostly come from different harnesses, effort levels and task subsets. Every benchmark.

Model Lab says Board says Gap
Inkling 2 Small 1,401 DesignArena 3D Design 1,460 Leaderboard −59

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