Benchmark Almanac

DesignArena Game Dev

GPT-6 Astra 1,470 leads since September 2026
In use

Models build a playable browser game as a single HTML file from a real user prompt. Users play the games 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
171
Entries on its own leaderboard, 143 models we track
1
Launch post lists it, from 1 lab

GPT-6 Astra leads with 1,470

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.

171 entries. Scroll sideways for the rest. Bars start at 600.

  1. 1,470 GPT-6 Astra
  2. 1,433 Inkling 2 Small
  3. 1,428 Claude Opus 5.5
  4. 1,416 Muse Wonder
  5. 1,407 GPT-6.1 Sol
  6. 1,406 Kimi K3.1 Pro
  7. 1,399 Claude Fable 5.2
  8. 1,391 Claude Sonnet 5.5
  9. 1,365 GPT-5.6 Sol
  10. 1,365 Claude Fable 5.1
  11. 1,362 Gemini 4 Argon
  12. 1,353 Kimi K3
  13. 1,346 Muse Spark 1.3
  14. 1,333 GPT-6.1 Terra
  15. 1,332 Muse Ember
  16. 1,331 Claude Opus 5
  17. 1,326 Claude Fable 5
  18. 1,325 DeepSeek V4.1 Flash
  19. 1,322 Gemini 4 Carbon
  20. 1,316 Claude Haiku 5.5
  21. 1,313 Gemini 3.7 Flash
  22. 1,313 MiMo-V2.6 Pro
  23. 1,311 Qwen 3.8 Max
  24. 1,311 MiniMax-M3.1-Flash
  25. 1,309 Ling-3.1-flash
  26. 1,307 Grok 4.7
  27. 1,307 GPT-6 Sol
  28. 1,297 Gemini 3.8 Flash
  29. 1,294 Muse Spark 1.2
  30. 1,290 GLM-5.3
  31. 1,283 GPT-5.5
  32. 1,282 Grok 4.6
  33. 1,272 Claude Sonnet 5
  34. 1,269 Tencent Hy4
  35. 1,268 Claude Opus 4.6
  36. 1,267 GLM-5.3-Flash
  37. 1,265 MiMo-V2.5-Pro
  38. 1,264 Muse Spark Open
  39. 1,258 Muse Spark 1.1
  40. 1,257 GLM-5.2
  41. 1,256 GLM-5.1
  42. 1,255 Qwen3.7-Max
  43. 1,254 Gemini 3.5 Flash
  44. 1,254 GPT-6 Luna
  45. 1,252 Deca 2.5 Ultra
  46. 1,252 Deca 2.6 Ultra
  47. 1,252 Claude Opus 4.8
  48. 1,252 Grok 4.5
  49. 1,252 Gemini 3.6 Flash
  50. 1,250 Qwen3.7-Plus
  51. 1,249 Claude Sonnet 4.6
  52. 1,246 Deca 2.5 Mini
  53. 1,243 Deca 2.6 Pro
  54. 1,243 MiMo-V2.5
  55. 1,240 Kimi K2.6
  56. 1,238 GLM-5-Turbo
  57. 1,229 GPT-5.4
  58. 1,224 GLM-5
  59. 1,222 DeepSeek-V4-Pro
  60. 1,220 GPT-5.6 Terra
  61. 1,219 Claude Opus 4.5
  62. 1,216 Nex N2 Pro
  63. 1,215 Muse Spark
  64. 1,210 Kimi K2.7 Code
  65. 1,207 GLM-5V-Turbo
  66. 1,205 Mistral Large 4
  67. 1,205 Mercury 3
  68. 1,202 Qwen 3.6 Plus
  69. 1,201 MiniMax-M3
  70. 1,199 MiniMax-M2.7
  71. 1,199 DeepSeek V4 Flash 0731
  72. 1,198 Tencent Hy3
  73. 1,197 Kimi K2.5
  74. 1,192 Gemini 3.1 Pro
  75. 1,191 GPT-5.6 Luna
  76. 1,187 DeepSeek-V4-Flash
  77. 1,186 Gemini 3 Pro
  78. 1,186 Beam 1.0
  79. 1,185 GPT-5.2
  80. 1,183 Grok 4.20 Beta
  81. 1,178 GLM-4.7
  82. 1,175 GPT-5
  83. 1,169 GPT-5.1
  84. 1,168 Solar Pro 4
  85. 1,167 AGI-01 Swift
  86. 1,166 MiniMax-M2.5
  87. 1,161 Claude Opus 4.1
  88. 1,161 Grok 4.3
  89. 1,160 Claude Opus 4
  90. 1,156 Claude Sonnet 4.5
  91. 1,156 Gemini 3 Flash
  92. 1,153 Inkling
  93. 1,151 GPT-5.3 Codex
  94. 1,144 Step 3.7 Flash
  95. 1,141 GLM-4.6
  96. 1,135 GLM-4.5
  97. 1,134 DeepSeek-V3.2-Exp
  98. 1,131 Deca 2.6 Mini
  99. 1,130 Qwen3.5 397B-A17B
  100. 1,129 Claude Sonnet 4
  101. 1,128 Claude 3.7 Sonnet
  102. 1,128 GPT-5.1-Codex
  103. 1,127 Nemotron 3 Ultra
  104. 1,125 glm-4.7-flash
  105. 1,124 MiniMax-M2.1
  106. 1,122 DeepSeek-V3.2
  107. 1,121 DeepSeek-V3.1
  108. 1,121 MiMo-V2-Flash
  109. 1,119 GPT-5 mini (Default)
  110. 1,118 v0-1.5-lg
  111. 1,111 Muse Glimmer 30B
  112. 1,108 MiniMax M2 Stable
  113. 1,101 Gemini 2.5 Pro
  114. 1,098 Solar Mini 4
  115. 1,095 Qwen3.5 Plus 02-15
  116. 1,089 Qwen3 Coder 480B A35B Instruct
  117. 1,087 DeepSeek-R1
  118. 1,087 Claude Haiku 4.5
  119. 1,084 GLM-4.5-Air
  120. 1,084 Qwen3-Max
  121. 1,082 GPT-5.1-Codex-mini
  122. 1,081 Flint
  123. 1,077 Nemotron 3.5 Super VL
  124. 1,076 Kimi K2 Turbo Preview
  125. 1,072 AesCoder-4B
  126. 1,070 GPT-4.1
  127. 1,070 Mistral Large 3
  128. 1,069 Trinity Large Thinking
  129. 1,069 Mistral Medium 3.5
  130. 1,067 Prime Intellect: INTELLECT-3
  131. 1,064 Mistral Medium 3.1
  132. 1,061 GPT-4.1 mini
  133. 1,057 Grok 4.1 Fast
  134. 1,056 Grok 4 Fast
  135. 1,056 Gemini 2.5 Flash
  136. 1,043 DeepSeek-V3
  137. 1,037 Gemini 2.5 Flash-Lite
  138. 1,035 GPT-5 nano (Default)
  139. 1,028 Ministral 3 14B
  140. 1,024 o3
  141. 1,020 Gemini 3.1 Flash-Lite
  142. 1,016 Grok 3
  143. 1,013 v0-1.5-md
  144. 1,008 Mistral Medium 3 (2505)
  145. 993 o4-mini
  146. 992 Grok 4
  147. 986 Magistral Medium 1.2 (2509)
  148. 982 gpt-oss-120b
  149. 980 Ministral 3 8B Instruct
  150. 978 grok-code-fast-1
  151. 976 Devstral Medium
  152. 964 Kimi K2
  153. 963 Mercury 2
  154. 961 GPT-4.1 nano
  155. 958 Codestral 2508
  156. 957 Olmo 3.1 32B Think
  157. 947 Qwen3-235B-A22B-Thinking
  158. 942 Qwen3-235B-A22B-Instruct-2507
  159. 940 Grok-3 mini
  160. 939 Ministral 3 3B Instruct
  161. 932 Magistral Small 1.2 (2509)
  162. 918 Qwen3-235B-A22B
  163. 896 GPT-4o
  164. 889 Qwen3-30B-A3B
  165. 878 Mistral Large 2.1 (2411)
  166. 878 Mistral Small 3.2
  167. 860 Llama 3.1 Nemotron Ultra 253B
  168. 828 Llama 4 Maverick
  169. 802 Codestral 2 (2501)
  170. 771 Devstral Small 1.1
  171. 763 Llama 4 Scout

Bars start at 600.

  1. 1 GPT-6 Astra xhigh, max 1,470
  2. 2 Inkling 2 Small 1,433
  3. 3 Claude Opus 5.5 1,428
  4. 4 Muse Wonder max 1,416
  5. 5 GPT-6.1 Sol 1,407
  6. 6 Kimi K3.1 Pro 1,406
  7. 7 Claude Fable 5.2 1,399
  8. 8 Claude Sonnet 5.5 1,391
  9. 9 GPT-5.6 Sol medium, xhigh 1,365
  10. 10 Claude Fable 5.1 1,365
  11. 11 Gemini 4 Argon 1,362
  12. 12 Kimi K3 1,353
  13. 13 Muse Spark 1.3 xhigh, max 1,346
  14. 14 GPT-6.1 Terra 1,333
  15. 15 Muse Ember 1,332
  16. 16 Claude Opus 5 1,331
  17. 17 Claude Fable 5 1,326
  18. 18 DeepSeek V4.1 Flash 1,325
  19. 19 Gemini 4 Carbon 1,322
  20. 20 Claude Haiku 5.5 1,316
  21. 21 Gemini 3.7 Flash 1,313
  22. 22 MiMo-V2.6 Pro 1,313
  23. 23 Qwen 3.8 Max 1,311
  24. 24 MiniMax-M3.1-Flash 1,311
  25. 25 Ling-3.1-flash 1,309
  26. 26 Grok 4.7 1,307
  27. 27 GPT-6 Sol 1,307
  28. 28 Gemini 3.8 Flash 1,297
  29. 29 Muse Spark 1.2 1,294
  30. 30 GLM-5.3 1,290
  31. 31 GPT-5.5 1,283
  32. 32 Grok 4.6 1,282
  33. 33 Claude Sonnet 5 1,272
  34. 34 Tencent Hy4 1,269
  35. 35 Claude Opus 4.6 thinking 1,268
  36. 36 GLM-5.3-Flash 1,267
  37. 37 MiMo-V2.5-Pro 1,265
  38. 38 Muse Spark Open xhigh 1,264
  39. 39 Muse Spark 1.1 1,258
  40. 40 GLM-5.2 1,257
  41. 41 GLM-5.1 1,256
  42. 42 Qwen3.7-Max 1,255
  43. 43 Gemini 3.5 Flash 1,254
  44. 44 GPT-6 Luna 1,254
  45. 45 Deca 2.5 Ultra 2 1,252
  46. 46 Deca 2.6 Ultra 1,252
  47. 47 Claude Opus 4.8 1,252
  48. 48 Grok 4.5 1,252
  49. 49 Gemini 3.6 Flash 1,252
  50. 50 Qwen3.7-Plus 1,250
  51. 51 Claude Sonnet 4.6 1,249
  52. 52 Deca 2.5 Mini 1,246
  53. 53 Deca 2.6 Pro 1,243
  54. 54 MiMo-V2.5 1,243
  55. 55 Kimi K2.6 1,240
  56. 56 GLM-5-Turbo 1,238
  57. 57 GPT-5.4 medium 1,229
  58. 58 GLM-5 1,224
  59. 59 DeepSeek-V4-Pro 1,222
  60. 60 GPT-5.6 Terra 1,220
  61. 61 Claude Opus 4.5 1,219
  62. 62 Nex N2 Pro 1,216
  63. 63 Muse Spark 1,215
  64. 64 Kimi K2.7 Code 1,210
  65. 65 GLM-5V-Turbo 1,207
  66. 66 Mistral Large 4 1,205
  67. 67 Mercury 3 1,205
  68. 68 Qwen 3.6 Plus 1,202
  69. 69 MiniMax-M3 1,201
  70. 70 MiniMax-M2.7 1,199
  71. 71 DeepSeek V4 Flash 0731 1,199
  72. 72 Tencent Hy3 1,198
  73. 73 Kimi K2.5 thinking 1,197
  74. 74 Gemini 3.1 Pro 1,192
  75. 75 GPT-5.6 Luna 1,191
  76. 76 DeepSeek-V4-Flash 1,187
  77. 77 Gemini 3 Pro 1,186
  78. 78 Beam 1.0 max 1,186
  79. 79 GPT-5.2 low 1,185
  80. 80 Grok 4.20 Beta reasoning 1,183
  81. 81 GLM-4.7 1,178
  82. 82 GPT-5 minimal, high 1,175
  83. 83 GPT-5.1 high 1,169
  84. 84 Solar Pro 4 1,168
  85. 85 AGI-01 Swift 1,167
  86. 86 MiniMax-M2.5 1,166
  87. 87 Claude Opus 4.1 thinking 1,161
  88. 88 Grok 4.3 1,161
  89. 89 Claude Opus 4 1,160
  90. 90 Claude Sonnet 4.5 1,156
  91. 91 Gemini 3 Flash 1,156
  92. 92 Inkling 1,153
  93. 93 GPT-5.3 Codex 1,151
  94. 94 Step 3.7 Flash 1,144
  95. 95 GLM-4.6 1,141
  96. 96 GLM-4.5 1,135
  97. 97 DeepSeek-V3.2-Exp 1,134
  98. 98 Deca 2.6 Mini 1,131
  99. 99 Qwen3.5 397B-A17B 1,130
  100. 100 Claude Sonnet 4 thinking 1,129
  101. 101 Claude 3.7 Sonnet 1,128
  102. 102 GPT-5.1-Codex 1,128
  103. 103 Nemotron 3 Ultra 1,127
  104. 104 glm-4.7-flash 1,125
  105. 105 MiniMax-M2.1 1,124
  106. 106 DeepSeek-V3.2 1,122
  107. 107 DeepSeek-V3.1 thinking 1,121
  108. 108 MiMo-V2-Flash 1,121
  109. 109 GPT-5 mini (Default) 1,119
  110. 110 v0-1.5-lg 1,118
  111. 111 Muse Glimmer 30B xhigh 1,111
  112. 112 MiniMax M2 Stable 1,108
  113. 113 Gemini 2.5 Pro Jun 2025 1,101
  114. 114 Solar Mini 4 1,098
  115. 115 Qwen3.5 Plus 02-15 1,095
  116. 116 Qwen3 Coder 480B A35B Instruct 1,089
  117. 117 DeepSeek-R1 May 2025 1,087
  118. 118 Claude Haiku 4.5 1,087
  119. 119 GLM-4.5-Air 1,084
  120. 120 Qwen3-Max 1,084
  121. 121 GPT-5.1-Codex-mini 1,082
  122. 122 Flint 1,081
  123. 123 Nemotron 3.5 Super VL 1,077
  124. 124 Kimi K2 Turbo Preview 1,076
  125. 125 AesCoder-4B 1,072
  126. 126 GPT-4.1 1,070
  127. 127 Mistral Large 3 1,070
  128. 128 Trinity Large Thinking 1,069
  129. 129 Mistral Medium 3.5 1,069
  130. 130 Prime Intellect: INTELLECT-3 1,067
  131. 131 Mistral Medium 3.1 1,064
  132. 132 GPT-4.1 mini 1,061
  133. 133 Grok 4.1 Fast 1,057
  134. 134 Grok 4 Fast 1,056
  135. 135 Gemini 2.5 Flash Sep 2025 1,056
  136. 136 DeepSeek-V3 Mar 2025 1,043
  137. 137 Gemini 2.5 Flash-Lite Sep 2025 1,037
  138. 138 GPT-5 nano (Default) 1,035
  139. 139 Ministral 3 14B 1,028
  140. 140 o3 1,024
  141. 141 Gemini 3.1 Flash-Lite 1,020
  142. 142 Grok 3 1,016
  143. 143 v0-1.5-md 1,013
  144. 144 Mistral Medium 3 (2505) 1,008
  145. 145 o4-mini 993
  146. 146 Grok 4 992
  147. 147 Magistral Medium 1.2 (2509) 986
  148. 148 gpt-oss-120b 982
  149. 149 Ministral 3 8B Instruct 980
  150. 150 grok-code-fast-1 978
  151. 151 Devstral Medium 976
  152. 152 Kimi K2 Sep 2025 964
  153. 153 Mercury 2 963
  154. 154 GPT-4.1 nano 961
  155. 155 Codestral 2508 958
  156. 156 Olmo 3.1 32B Think 957
  157. 157 Qwen3-235B-A22B-Thinking Jul 2025 947
  158. 158 Qwen3-235B-A22B-Instruct-2507 942
  159. 159 Grok-3 mini 940
  160. 160 Ministral 3 3B Instruct 939
  161. 161 Magistral Small 1.2 (2509) 932
  162. 162 Qwen3-235B-A22B 918
  163. 163 GPT-4o Mar 2025 896
  164. 164 Qwen3-30B-A3B 889
  165. 165 Mistral Large 2.1 (2411) 878
  166. 166 Mistral Small 3.2 878
  167. 167 Llama 3.1 Nemotron Ultra 253B 860
  168. 168 Llama 4 Maverick 828
  169. 169 Codestral 2 (2501) 802
  170. 170 Devstral Small 1.1 771
  171. 171 Llama 4 Scout 763

Source: Official DesignArena Game Dev leaderboard

#EntryScoreSettingDateSource
1 GPT-6 Astra 1,470 xhigh, max Sep 2026 Official leaderboard
2 Inkling 2 Small 1,433 Nov 2026 Official leaderboard
3 Claude Opus 5.5 1,428 Sep 2026 Official leaderboard
4 Muse Wonder 1,416 max Oct 2026 Official leaderboard
5 GPT-6.1 Sol 1,407 Sep 2026 Official leaderboard
6 Kimi K3.1 Pro 1,406 Oct 2026 Official leaderboard
7 Claude Fable 5.2 1,399 Nov 2026 Official leaderboard
8 Claude Sonnet 5.5 1,391 Sep 2026 Official leaderboard
9 GPT-5.6 Sol 1,365 medium, xhigh Jul 2026 Official leaderboard
10 Claude Fable 5.1 1,365 Sep 2026 Official leaderboard
11 Gemini 4 Argon 1,362 Sep 2026 Official leaderboard
12 Kimi K3 1,353 Jul 2026 Official leaderboard
13 Muse Spark 1.3 1,346 xhigh, max Sep 2026 Official leaderboard
14 GPT-6.1 Terra 1,333 Oct 2026 Official leaderboard
15 Muse Ember 1,332 Nov 2026 Official leaderboard
16 Claude Opus 5 1,331 Jul 2026 Official leaderboard
17 Claude Fable 5 1,326 Jun 2026 Official leaderboard
18 DeepSeek V4.1 Flash 1,325 Sep 2026 Official leaderboard
19 Gemini 4 Carbon 1,322 Oct 2026 Official leaderboard
20 Claude Haiku 5.5 1,316 Oct 2026 Official leaderboard
21 Gemini 3.7 Flash 1,313 Aug 2026 Official leaderboard
22 MiMo-V2.6 Pro 1,313 Sep 2026 Official leaderboard
23 Qwen 3.8 Max 1,311 Aug 2026 Official leaderboard
24 MiniMax-M3.1-Flash 1,311 Oct 2026 Official leaderboard
25 Ling-3.1-flash 1,309 Oct 2026 Official leaderboard
26 Grok 4.7 1,307 Sep 2026 Official leaderboard
27 GPT-6 Sol 1,307 Sep 2026 Official leaderboard
28 Gemini 3.8 Flash 1,297 Sep 2026 Official leaderboard
29 Muse Spark 1.2 1,294 Aug 2026 Official leaderboard
30 GLM-5.3 1,290 Aug 2026 Official leaderboard
31 GPT-5.5 1,283 Apr 2026 Official leaderboard
32 Grok 4.6 1,282 Aug 2026 Official leaderboard
33 Claude Sonnet 5 1,272 Jun 2026 Official leaderboard
34 Tencent Hy4 1,269 Nov 2026 Official leaderboard
35 Claude Opus 4.6 1,268 thinking Feb 2026 Official leaderboard
36 GLM-5.3-Flash 1,267 Aug 2026 Official leaderboard
37 MiMo-V2.5-Pro 1,265 Apr 2026 Official leaderboard
38 Muse Spark Open 1,264 xhigh Oct 2026 Official leaderboard
39 Muse Spark 1.1 1,258 Jul 2026 Official leaderboard
40 GLM-5.2 1,257 Jun 2026 Official leaderboard
41 GLM-5.1 1,256 Apr 2026 Official leaderboard
42 Qwen3.7-Max 1,255 May 2026 Official leaderboard
43 Gemini 3.5 Flash 1,254 May 2026 Official leaderboard
44 GPT-6 Luna 1,254 Sep 2026 Official leaderboard
45 Deca 2.5 Ultra 1,252 2 Official leaderboard
46 Deca 2.6 Ultra 1,252 Official leaderboard
47 Claude Opus 4.8 1,252 May 2026 Official leaderboard
48 Grok 4.5 1,252 Jul 2026 Official leaderboard
49 Gemini 3.6 Flash 1,252 Jul 2026 Official leaderboard
50 Qwen3.7-Plus 1,250 Jun 2026 Official leaderboard
51 Claude Sonnet 4.6 1,249 Feb 2026 Official leaderboard
52 Deca 2.5 Mini 1,246 Official leaderboard
53 Deca 2.6 Pro 1,243 Official leaderboard
54 MiMo-V2.5 1,243 Apr 2026 Official leaderboard
55 Kimi K2.6 1,240 Apr 2026 Official leaderboard
56 GLM-5-Turbo 1,238 Mar 2026 Official leaderboard
57 GPT-5.4 1,229 medium Mar 2026 Official leaderboard
58 GLM-5 1,224 Feb 2026 Official leaderboard
59 DeepSeek-V4-Pro 1,222 Apr 2026 Official leaderboard
60 GPT-5.6 Terra 1,220 Jul 2026 Official leaderboard
61 Claude Opus 4.5 1,219 Nov 2025 Official leaderboard
62 Nex N2 Pro 1,216 Official leaderboard
63 Muse Spark 1,215 Apr 2026 Official leaderboard
64 Kimi K2.7 Code 1,210 Jun 2026 Official leaderboard
65 GLM-5V-Turbo 1,207 Apr 2026 Official leaderboard
66 Mistral Large 4 1,205 Oct 2026 Official leaderboard
67 Mercury 3 1,205 Oct 2026 Official leaderboard
68 Qwen 3.6 Plus 1,202 Mar 2026 Official leaderboard
69 MiniMax-M3 1,201 Jun 2026 Official leaderboard
70 MiniMax-M2.7 1,199 Mar 2026 Official leaderboard
71 DeepSeek V4 Flash 0731 1,199 Jul 2026 Official leaderboard
72 Tencent Hy3 1,198 Jul 2026 Official leaderboard
73 Kimi K2.5 1,197 thinking Jan 2026 Official leaderboard
74 Gemini 3.1 Pro 1,192 Feb 2026 Official leaderboard
75 GPT-5.6 Luna 1,191 Jul 2026 Official leaderboard
76 DeepSeek-V4-Flash 1,187 Apr 2026 Official leaderboard
77 Gemini 3 Pro 1,186 Nov 2025 Official leaderboard
78 Beam 1.0 1,186 max Oct 2026 Official leaderboard
79 GPT-5.2 1,185 low Dec 2025 Official leaderboard
80 Grok 4.20 Beta 1,183 reasoning Official leaderboard
81 GLM-4.7 1,178 Dec 2025 Official leaderboard
82 GPT-5 1,175 minimal, high Aug 2025 Official leaderboard
83 GPT-5.1 1,169 high Nov 2025 Official leaderboard
84 Solar Pro 4 1,168 Aug 2026 Official leaderboard
85 AGI-01 Swift 1,167 Official leaderboard
86 MiniMax-M2.5 1,166 Feb 2026 Official leaderboard
87 Claude Opus 4.1 1,161 thinking Aug 2025 Official leaderboard
88 Grok 4.3 1,161 Apr 2026 Official leaderboard
89 Claude Opus 4 1,160 May 2025 Official leaderboard
90 Claude Sonnet 4.5 1,156 Sep 2025 Official leaderboard
91 Gemini 3 Flash 1,156 Dec 2025 Official leaderboard
92 Inkling 1,153 Jul 2026 Official leaderboard
93 GPT-5.3 Codex 1,151 Feb 2026 Official leaderboard
94 Step 3.7 Flash 1,144 May 2026 Official leaderboard
95 GLM-4.6 1,141 Sep 2025 Official leaderboard
96 GLM-4.5 1,135 Aug 2025 Official leaderboard
97 DeepSeek-V3.2-Exp 1,134 Sep 2025 Official leaderboard
98 Deca 2.6 Mini 1,131 Official leaderboard
99 Qwen3.5 397B-A17B 1,130 Feb 2026 Official leaderboard
100 Claude Sonnet 4 1,129 thinking May 2025 Official leaderboard
101 Claude 3.7 Sonnet 1,128 Feb 2025 Official leaderboard
102 GPT-5.1-Codex 1,128 Nov 2025 Official leaderboard
103 Nemotron 3 Ultra 1,127 Jun 2026 Official leaderboard
104 glm-4.7-flash 1,125 Jan 2026 Official leaderboard
105 MiniMax-M2.1 1,124 Dec 2025 Official leaderboard
106 DeepSeek-V3.2 1,122 Dec 2025 Official leaderboard
107 DeepSeek-V3.1 1,121 thinking Aug 2025 Official leaderboard
108 MiMo-V2-Flash 1,121 Dec 2025 Official leaderboard
109 GPT-5 mini (Default) 1,119 Official leaderboard
110 v0-1.5-lg 1,118 Official leaderboard
111 Muse Glimmer 30B 1,111 xhigh Official leaderboard
112 MiniMax M2 Stable 1,108 Official leaderboard
113 Gemini 2.5 Pro Jun 2025 1,101 Jun 2025 Official leaderboard
114 Solar Mini 4 1,098 Oct 2026 Official leaderboard
115 Qwen3.5 Plus 02-15 1,095 Official leaderboard
116 Qwen3 Coder 480B A35B Instruct 1,089 Official leaderboard
117 DeepSeek-R1 May 2025 1,087 May 2025 Official leaderboard
118 Claude Haiku 4.5 1,087 Oct 2025 Official leaderboard
119 GLM-4.5-Air 1,084 Jul 2025 Official leaderboard
120 Qwen3-Max 1,084 Sep 2025 Official leaderboard
121 GPT-5.1-Codex-mini 1,082 Nov 2025 Official leaderboard
122 Flint 1,081 Official leaderboard
123 Nemotron 3.5 Super VL 1,077 Oct 2026 Official leaderboard
124 Kimi K2 Turbo Preview 1,076 Official leaderboard
125 AesCoder-4B 1,072 Official leaderboard
126 GPT-4.1 1,070 Apr 2025 Official leaderboard
127 Mistral Large 3 1,070 Dec 2025 Official leaderboard
128 Trinity Large Thinking 1,069 Apr 2026 Official leaderboard
129 Mistral Medium 3.5 1,069 Apr 2026 Official leaderboard
130 Prime Intellect: INTELLECT-3 1,067 Official leaderboard
131 Mistral Medium 3.1 1,064 Aug 2025 Official leaderboard
132 GPT-4.1 mini 1,061 Apr 2025 Official leaderboard
133 Grok 4.1 Fast 1,057 Nov 2025 Official leaderboard
134 Grok 4 Fast 1,056 Sep 2025 Official leaderboard
135 Gemini 2.5 Flash Sep 2025 1,056 Sep 2025 Official leaderboard
136 DeepSeek-V3 Mar 2025 1,043 Mar 2025 Official leaderboard
137 Gemini 2.5 Flash-Lite Sep 2025 1,037 Sep 2025 Official leaderboard
138 GPT-5 nano (Default) 1,035 Official leaderboard
139 Ministral 3 14B 1,028 Dec 2025 Official leaderboard
140 o3 1,024 Apr 2025 Official leaderboard
141 Gemini 3.1 Flash-Lite 1,020 Mar 2026 Official leaderboard
142 Grok 3 1,016 Apr 2025 Official leaderboard
143 v0-1.5-md 1,013 Official leaderboard
144 Mistral Medium 3 (2505) 1,008 Official leaderboard
145 o4-mini 993 Apr 2025 Official leaderboard
146 Grok 4 992 Jul 2025 Official leaderboard
147 Magistral Medium 1.2 (2509) 986 Official leaderboard
148 gpt-oss-120b 982 Aug 2025 Official leaderboard
149 Ministral 3 8B Instruct 980 Dec 2025 Official leaderboard
150 grok-code-fast-1 978 Aug 2025 Official leaderboard
151 Devstral Medium 976 Jul 2025 Official leaderboard
152 Kimi K2 Sep 2025 964 Sep 2025 Official leaderboard
153 Mercury 2 963 Feb 2026 Official leaderboard
154 GPT-4.1 nano 961 Apr 2025 Official leaderboard
155 Codestral 2508 958 Official leaderboard
156 Olmo 3.1 32B Think 957 Dec 2025 Official leaderboard
157 Qwen3-235B-A22B-Thinking Jul 2025 947 Jul 2025 Official leaderboard
158 Qwen3-235B-A22B-Instruct-2507 942 Official leaderboard
159 Grok-3 mini 940 Apr 2025 Official leaderboard
160 Ministral 3 3B Instruct 939 Dec 2025 Official leaderboard
161 Magistral Small 1.2 (2509) 932 Official leaderboard
162 Qwen3-235B-A22B 918 Apr 2025 Official leaderboard
163 GPT-4o Mar 2025 896 Mar 2025 Official leaderboard
164 Qwen3-30B-A3B 889 Apr 2025 Official leaderboard
165 Mistral Large 2.1 (2411) 878 Official leaderboard
166 Mistral Small 3.2 878 Jun 2025 Official leaderboard
167 Llama 3.1 Nemotron Ultra 253B 860 Official leaderboard
168 Llama 4 Maverick 828 Apr 2025 Official leaderboard
169 Codestral 2 (2501) 802 Official leaderboard
170 Devstral Small 1.1 771 Jul 2025 Official leaderboard
171 Llama 4 Scout 763 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 30 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
1,470
GPT-5.6 Sol
medium medium, xhi…
1,365
Muse Spark 1.3
xhigh
1,346
DeepSeek V4.1 Flash
1017
1,325
Claude Opus 4.6
thinking
1,268
GPT-5.4
none medium
1,229
Tencent Hy3
low
1,198
GPT-5.2
none
1,185
GPT-5
minimal minimal, hi…
1,175
GPT-5.1
none high
1,169
Claude Opus 4.1
thinking
1,161
Claude Sonnet 4.5
thinking
1,156

Source: Official DesignArena Game Dev leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-6 Astra xhigh, max 1,470
GPT-6 Astra xhigh 1,460
Inkling 2 Small 1,433
Claude Opus 5.5 1,428
Muse Wonder max 1,416
GPT-6.1 Sol 1,407
Kimi K3.1 Pro 1,406
Claude Fable 5.2 1,399
Claude Sonnet 5.5 1,391
GPT-5.6 Sol medium, xhigh 1,365
Claude Fable 5.1 1,365
Gemini 4 Argon 1,362
Kimi K3 1,353
Muse Spark 1.3 xhigh, max 1,346
GPT-6.1 Terra 1,333
Muse Ember 1,332
Claude Opus 5 1,331
Muse Spark 1.3 xhigh 1,331
Claude Fable 5 1,326
DeepSeek V4.1 Flash 1,325
Gemini 4 Carbon 1,322
GPT-5.6 Sol medium 1,317
Claude Haiku 5.5 1,316
Gemini 3.7 Flash 1,313
MiMo-V2.6 Pro 1,313
Qwen 3.8 Max 1,311
MiniMax-M3.1-Flash 1,311
Ling-3.1-flash 1,309
Grok 4.7 1,307
GPT-6 Sol 1,307
Gemini 3.8 Flash 1,297
Muse Spark 1.2 1,294
GLM-5.3 1,290
GPT-5.5 1,283
Grok 4.6 1,282
Claude Sonnet 5 1,272
Tencent Hy4 1,269
Claude Opus 4.6 thinking 1,268
GLM-5.3-Flash 1,267
MiMo-V2.5-Pro 1,265
Muse Spark Open xhigh 1,264
Claude Opus 4.6 1,262
Muse Spark 1.1 1,258
GLM-5.2 1,257
GLM-5.1 1,256
Qwen3.7-Max 1,255
Gemini 3.5 Flash 1,254
GPT-6 Luna 1,254
Claude Opus 4.8 1,252
Grok 4.5 1,252
Gemini 3.6 Flash 1,252
DeepSeek V4.1 Flash 1017 1,251
Qwen3.7-Plus 1,250
Claude Sonnet 4.6 1,249
MiMo-V2.5 1,243
Kimi K2.6 1,240
GLM-5-Turbo 1,238
GPT-5.4 medium 1,229
GLM-5 1,224
DeepSeek-V4-Pro 1,222
GPT-5.6 Terra 1,220
Claude Opus 4.5 1,219
Muse Spark 1,215
Kimi K2.7 Code 1,210
GLM-5V-Turbo 1,207
Mistral Large 4 1,205
Mercury 3 1,205
Qwen 3.6 Plus 1,202
MiniMax-M3 1,201
MiniMax-M2.7 1,199
DeepSeek V4 Flash 0731 1,199
Tencent Hy3 1,198
Kimi K2.5 thinking 1,197
Gemini 3.1 Pro 1,192
GPT-5.6 Luna 1,191
DeepSeek-V4-Flash 1,187
Gemini 3 Pro 1,186
Beam 1.0 max 1,186
GPT-5.2 low 1,185
GPT-5.4 none 1,184
GPT-5.4 low 1,181
GLM-4.7 1,178
GPT-5.2 high 1,177
GPT-5 minimal, high 1,175
GPT-5.2 medium 1,174
GPT-5.1 high 1,169
Solar Pro 4 1,168
GPT-5.2 none 1,166
MiniMax-M2.5 1,166
Claude Opus 4.1 thinking 1,161
Grok 4.3 1,161
Claude Opus 4 1,160
Claude Sonnet 4.5 1,156
Gemini 3 Flash 1,156
Claude Opus 4.1 1,153
Claude Sonnet 4.5 thinking 1,153
Inkling 1,153
GPT-5.3 Codex 1,151
GPT-5.1 medium 1,149
GPT-5 minimal 1,145
Step 3.7 Flash 1,144
GLM-4.6 1,141
GPT-5.1 low 1,141
GLM-4.5 1,135
DeepSeek-V3.2-Exp 1,134
Qwen3.5 397B-A17B 1,130
Claude Sonnet 4 thinking 1,129
Claude 3.7 Sonnet 1,128
GPT-5.1-Codex 1,128
Nemotron 3 Ultra 1,127
Tencent Hy3 low 1,126
glm-4.7-flash 1,125
MiniMax-M2.1 1,124
Claude Sonnet 4 1,123
GPT-5.1 none 1,122
DeepSeek-V3.2 1,122
DeepSeek-V3.1 thinking 1,121
MiMo-V2-Flash 1,121
Gemini 2.5 Pro Jun 2025 1,101
Solar Mini 4 1,098
DeepSeek-R1 May 2025 1,087
Claude Haiku 4.5 1,087
GLM-4.5-Air 1,084
Qwen3-Max 1,084
GPT-5.1-Codex-mini 1,082
Nemotron 3.5 Super VL 1,077
DeepSeek-V3.1 1,074
GPT-4.1 1,070
Mistral Large 3 1,070
Trinity Large Thinking 1,069
Mistral Medium 3.5 1,069
Mistral Medium 3.1 1,064
GPT-4.1 mini 1,061
Grok 4.1 Fast 1,057
Grok 4 Fast 1,056
Gemini 2.5 Flash Sep 2025 1,056
Grok 4.1 Fast reasoning 1,050
DeepSeek-V3 Mar 2025 1,043
Gemini 2.5 Flash-Lite Sep 2025 1,037
Ministral 3 14B 1,028
o3 1,024
Gemini 3.1 Flash-Lite 1,020
Grok 3 1,016
Grok 4 Fast reasoning 1,012
Gemini 2.5 Flash Sep 2025 995
o4-mini 993
Grok 4 992
gpt-oss-120b 982
Ministral 3 8B Instruct 980
grok-code-fast-1 978
Devstral Medium 976
Kimi K2 Sep 2025 964
Mercury 2 963
GPT-4.1 nano 961
Olmo 3.1 32B Think 957
Qwen3-235B-A22B-Thinking Jul 2025 947
Grok-3 mini 940
Ministral 3 3B Instruct 939
Qwen3-235B-A22B 918
GPT-4o Mar 2025 896
Qwen3-30B-A3B 889
Mistral Small 3.2 878
Llama 4 Maverick 828
Devstral Small 1.1 771
Llama 4 Scout 763

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
Own leaderboard 11 leaders
GPT-6 Astra 1,470

Source: Official DesignArena Game Dev 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
600 800 1,000 1,200 1,400 1,600 2025 2026 2027 600 800 1,000 1,200 1,400 1,600 2025 2026 2027

Sources: Official DesignArena Game Dev 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 1428 DesignArena Game Dev

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 1,428 DesignArena Game Dev 1,433 Leaderboard −5

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