Benchmark Almanac

DesignArena Agentic Overall

Claude Opus 5.5 1,442 leads since September 2026
Not in model cards

Coding agents build apps, games, slides and spreadsheets from user prompts. Users compare the results, and votes from every agent board form one rating.

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

Kind
Building whole projects
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 model added
November 2026
Scores refreshed
November 12th, 2026
75
Entries on its own leaderboard, 74 models we track

Claude Opus 5.5 leads with 1,442

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.

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

  1. 1,442 Claude Opus 5.5
  2. 1,400 GPT-6.1 Sol
  3. 1,385 GPT-6 Astra
  4. 1,354 Claude Fable 5.2
  5. 1,327 Claude Fable 5.1
  6. 1,307 Gemini 4 Carbon
  7. 1,300 DeepSeek V4.1 Flash
  8. 1,283 Kimi K3.1 Pro
  9. 1,280 Claude Sonnet 5.5
  10. 1,275 Muse Spark 1.3
  11. 1,269 Inkling 2 Small
  12. 1,266 Claude Fable 5
  13. 1,264 Claude Opus 5
  14. 1,264 Gemini 4 Argon
  15. 1,263 MiMo-V2.6 Pro
  16. 1,254 Kimi K3
  17. 1,252 Qwen 3.8 Max
  18. 1,243 Grok 4.7
  19. 1,243 MiniMax-M3.1-Flash
  20. 1,233 GPT-5.6 Sol
  21. 1,230 Grok 4.6
  22. 1,226 GLM-5.3
  23. 1,225 Ling 3.2 Flash
  24. 1,218 Claude Opus 4.8
  25. 1,217 Gemini 3.8 Flash
  26. 1,216 Gemini 3.7 Flash
  27. 1,215 Claude Sonnet 5
  28. 1,213 Grok 4.5
  29. 1,209 GPT-6 Sol
  30. 1,198 Ling-3.1-flash
  31. 1,193 GPT-6.1 Terra
  32. 1,190 Claude Opus 4.6
  33. 1,187 GLM-5.2
  34. 1,180 MiniMax-M3
  35. 1,179 Claude Sonnet 4.6
  36. 1,178 Muse Spark 1.2
  37. 1,171 Qwen3.7-Max
  38. 1,169 Gemini 3.6 Flash
  39. 1,166 Gemini 3.5 Flash
  40. 1,165 Muse Spark 1.1
  41. 1,165 Inkling
  42. 1,162 Muse Wonder
  43. 1,154 GLM-5.1
  44. 1,151 Kimi K2.6
  45. 1,151 Kimi K2.7 Code
  46. 1,149 Claude Opus 4.5
  47. 1,146 GPT-6 Luna
  48. 1,141 GPT-5.6 Luna
  49. 1,109 Kimi K2.5
  50. 1,108 GLM-5
  51. 1,106 GPT-5.5
  52. 1,103 GLM-5V-Turbo
  53. 1,086 Grok 4.20
  54. 1,083 GPT-5.6 Terra
  55. 1,080 Gemini 3.1 Pro
  56. 1,076 Solar Pro 4
  57. 1,074 Gemini 3 Flash
  58. 1,073 Gemini 3 Pro
  59. 1,059 Claude Sonnet 4.5
  60. 1,057 GPT-5.2
  61. 1,055 GLM-4.7
  62. 1,052 GLM-4.6
  63. 1,043 GPT-5-Codex
  64. 1,035 Grok 4.3
  65. 1,034 GPT-5.4
  66. 1,032 GPT-5.2 Codex
  67. 1,027 GPT-5.1-Codex
  68. 1,010 GPT-5.3 Codex
  69. 1,007 grok-code-fast-1
  70. 1,001 GPT-5.1
  71. 933 Grok 4.1 Fast
  72. 919 DeepSeek-V4-Pro
  73. 908 Mistral Large
  74. 907 Gemini 2.5 Pro
  75. 887 Mistral Medium 3.5

Bars start at 800.

  1. 1 Claude Opus 5.5 high 1,442
  2. 2 GPT-6.1 Sol max 1,400
  3. 3 GPT-6 Astra xhigh, max 1,385
  4. 4 Claude Fable 5.2 1,354
  5. 5 Claude Fable 5.1 1,327
  6. 6 Gemini 4 Carbon 1,307
  7. 7 DeepSeek V4.1 Flash 1017 1,300
  8. 8 Kimi K3.1 Pro 1,283
  9. 9 Claude Sonnet 5.5 1,280
  10. 10 Muse Spark 1.3 xhigh, max 1,275
  11. 11 Inkling 2 Small 1,269
  12. 12 Claude Fable 5 1,266
  13. 13 Claude Opus 5 1,264
  14. 14 Gemini 4 Argon 1,264
  15. 15 MiMo-V2.6 Pro 1,263
  16. 16 Kimi K3 1,254
  17. 17 Qwen 3.8 Max 1,252
  18. 18 Grok 4.7 1,243
  19. 19 MiniMax-M3.1-Flash 1,243
  20. 20 GPT-5.6 Sol medium, xhigh 1,233
  21. 21 Grok 4.6 1,230
  22. 22 GLM-5.3 1,226
  23. 23 Ling 3.2 Flash 1,225
  24. 24 Claude Opus 4.8 1,218
  25. 25 Gemini 3.8 Flash 1,217
  26. 26 Gemini 3.7 Flash 1,216
  27. 27 Claude Sonnet 5 1,215
  28. 28 Grok 4.5 1,213
  29. 29 GPT-6 Sol 1,209
  30. 30 Ling-3.1-flash 1,198
  31. 31 GPT-6.1 Terra 1,193
  32. 32 Claude Opus 4.6 1,190
  33. 33 GLM-5.2 1,187
  34. 34 MiniMax-M3 1,180
  35. 35 Claude Sonnet 4.6 1,179
  36. 36 Muse Spark 1.2 1,178
  37. 37 Qwen3.7-Max 1,171
  38. 38 Gemini 3.6 Flash 1,169
  39. 39 Gemini 3.5 Flash 1,166
  40. 40 Muse Spark 1.1 1,165
  41. 41 Inkling 1,165
  42. 42 Muse Wonder max 1,162
  43. 43 GLM-5.1 1,154
  44. 44 Kimi K2.6 1,151
  45. 45 Kimi K2.7 Code 1,151
  46. 46 Claude Opus 4.5 1,149
  47. 47 GPT-6 Luna 1,146
  48. 48 GPT-5.6 Luna 1,141
  49. 49 Kimi K2.5 thinking 1,109
  50. 50 GLM-5 1,108
  51. 51 GPT-5.5 1,106
  52. 52 GLM-5V-Turbo 1,103
  53. 53 Grok 4.20 reasoning 1,086
  54. 54 GPT-5.6 Terra 1,083
  55. 55 Gemini 3.1 Pro 1,080
  56. 56 Solar Pro 4 1,076
  57. 57 Gemini 3 Flash 1,074
  58. 58 Gemini 3 Pro 1,073
  59. 59 Claude Sonnet 4.5 1,059
  60. 60 GPT-5.2 1,057
  61. 61 GLM-4.7 1,055
  62. 62 GLM-4.6 1,052
  63. 63 GPT-5-Codex 1,043
  64. 64 Grok 4.3 1,035
  65. 65 GPT-5.4 1,034
  66. 66 GPT-5.2 Codex 1,032
  67. 67 GPT-5.1-Codex 1,027
  68. 68 GPT-5.3 Codex 1,010
  69. 69 grok-code-fast-1 1,007
  70. 70 GPT-5.1 1,001
  71. 71 Grok 4.1 Fast reasoning 933
  72. 72 DeepSeek-V4-Pro 919
  73. 73 Mistral Large 908
  74. 74 Gemini 2.5 Pro Jun 2025 907
  75. 75 Mistral Medium 3.5 887

Source: Official DesignArena Agentic Overall leaderboard

#EntryScoreSettingDateSource
1 Claude Opus 5.5 1,442 high Sep 2026 Official leaderboard
2 GPT-6.1 Sol 1,400 max Sep 2026 Official leaderboard
3 GPT-6 Astra 1,385 xhigh, max Sep 2026 Official leaderboard
4 Claude Fable 5.2 1,354 Nov 2026 Official leaderboard
5 Claude Fable 5.1 1,327 Sep 2026 Official leaderboard
6 Gemini 4 Carbon 1,307 Oct 2026 Official leaderboard
7 DeepSeek V4.1 Flash 1,300 1017 Sep 2026 Official leaderboard
8 Kimi K3.1 Pro 1,283 Oct 2026 Official leaderboard
9 Claude Sonnet 5.5 1,280 Sep 2026 Official leaderboard
10 Muse Spark 1.3 1,275 xhigh, max Sep 2026 Official leaderboard
11 Inkling 2 Small 1,269 Nov 2026 Official leaderboard
12 Claude Fable 5 1,266 Jun 2026 Official leaderboard
13 Claude Opus 5 1,264 Jul 2026 Official leaderboard
14 Gemini 4 Argon 1,264 Sep 2026 Official leaderboard
15 MiMo-V2.6 Pro 1,263 Sep 2026 Official leaderboard
16 Kimi K3 1,254 Jul 2026 Official leaderboard
17 Qwen 3.8 Max 1,252 Aug 2026 Official leaderboard
18 Grok 4.7 1,243 Sep 2026 Official leaderboard
19 MiniMax-M3.1-Flash 1,243 Oct 2026 Official leaderboard
20 GPT-5.6 Sol 1,233 medium, xhigh Jul 2026 Official leaderboard
21 Grok 4.6 1,230 Aug 2026 Official leaderboard
22 GLM-5.3 1,226 Aug 2026 Official leaderboard
23 Ling 3.2 Flash 1,225 Official leaderboard
24 Claude Opus 4.8 1,218 May 2026 Official leaderboard
25 Gemini 3.8 Flash 1,217 Sep 2026 Official leaderboard
26 Gemini 3.7 Flash 1,216 Aug 2026 Official leaderboard
27 Claude Sonnet 5 1,215 Jun 2026 Official leaderboard
28 Grok 4.5 1,213 Jul 2026 Official leaderboard
29 GPT-6 Sol 1,209 Sep 2026 Official leaderboard
30 Ling-3.1-flash 1,198 Oct 2026 Official leaderboard
31 GPT-6.1 Terra 1,193 Oct 2026 Official leaderboard
32 Claude Opus 4.6 1,190 Feb 2026 Official leaderboard
33 GLM-5.2 1,187 Jun 2026 Official leaderboard
34 MiniMax-M3 1,180 Jun 2026 Official leaderboard
35 Claude Sonnet 4.6 1,179 Feb 2026 Official leaderboard
36 Muse Spark 1.2 1,178 Aug 2026 Official leaderboard
37 Qwen3.7-Max 1,171 May 2026 Official leaderboard
38 Gemini 3.6 Flash 1,169 Jul 2026 Official leaderboard
39 Gemini 3.5 Flash 1,166 May 2026 Official leaderboard
40 Muse Spark 1.1 1,165 Jul 2026 Official leaderboard
41 Inkling 1,165 Jul 2026 Official leaderboard
42 Muse Wonder 1,162 max Oct 2026 Official leaderboard
43 GLM-5.1 1,154 Apr 2026 Official leaderboard
44 Kimi K2.6 1,151 Apr 2026 Official leaderboard
45 Kimi K2.7 Code 1,151 Jun 2026 Official leaderboard
46 Claude Opus 4.5 1,149 Nov 2025 Official leaderboard
47 GPT-6 Luna 1,146 Sep 2026 Official leaderboard
48 GPT-5.6 Luna 1,141 Jul 2026 Official leaderboard
49 Kimi K2.5 1,109 thinking Jan 2026 Official leaderboard
50 GLM-5 1,108 Feb 2026 Official leaderboard
51 GPT-5.5 1,106 Apr 2026 Official leaderboard
52 GLM-5V-Turbo 1,103 Apr 2026 Official leaderboard
53 Grok 4.20 1,086 reasoning Feb 2026 Official leaderboard
54 GPT-5.6 Terra 1,083 Jul 2026 Official leaderboard
55 Gemini 3.1 Pro 1,080 Feb 2026 Official leaderboard
56 Solar Pro 4 1,076 Aug 2026 Official leaderboard
57 Gemini 3 Flash 1,074 Dec 2025 Official leaderboard
58 Gemini 3 Pro 1,073 Nov 2025 Official leaderboard
59 Claude Sonnet 4.5 1,059 Sep 2025 Official leaderboard
60 GPT-5.2 1,057 Dec 2025 Official leaderboard
61 GLM-4.7 1,055 Dec 2025 Official leaderboard
62 GLM-4.6 1,052 Sep 2025 Official leaderboard
63 GPT-5-Codex 1,043 Sep 2025 Official leaderboard
64 Grok 4.3 1,035 Apr 2026 Official leaderboard
65 GPT-5.4 1,034 Mar 2026 Official leaderboard
66 GPT-5.2 Codex 1,032 Dec 2025 Official leaderboard
67 GPT-5.1-Codex 1,027 Nov 2025 Official leaderboard
68 GPT-5.3 Codex 1,010 Feb 2026 Official leaderboard
69 grok-code-fast-1 1,007 Aug 2025 Official leaderboard
70 GPT-5.1 1,001 Nov 2025 Official leaderboard
71 Grok 4.1 Fast 933 reasoning Nov 2025 Official leaderboard
72 DeepSeek-V4-Pro 919 Apr 2026 Official leaderboard
73 Mistral Large 908 Feb 2024 Official leaderboard
74 Gemini 2.5 Pro Jun 2025 907 Jun 2025 Official leaderboard
75 Mistral Medium 3.5 887 Apr 2026 Official leaderboard

More effort adds a median 48 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
Claude Opus 5.5
high
1,442
GPT-6.1 Sol
max
1,400
GPT-6 Astra
xhigh xhigh, max
1,385
Muse Spark 1.3
xhigh
1,275
GPT-5.6 Sol
medium medium, xhi…
1,233

Source: Official DesignArena Agentic Overall leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Claude Opus 5.5 high 1,442
GPT-6.1 Sol max 1,400
GPT-6 Astra xhigh, max 1,385
Claude Fable 5.2 1,354
Claude Opus 5.5 1,331
Claude Fable 5.1 1,327
Gemini 4 Carbon 1,307
DeepSeek V4.1 Flash 1017 1,300
GPT-6 Astra xhigh 1,298
Kimi K3.1 Pro 1,283
Claude Sonnet 5.5 1,280
Muse Spark 1.3 xhigh, max 1,275
Inkling 2 Small 1,269
Claude Fable 5 1,266
Claude Opus 5 1,264
Gemini 4 Argon 1,264
MiMo-V2.6 Pro 1,263
Muse Spark 1.3 xhigh 1,256
Kimi K3 1,254
Qwen 3.8 Max 1,252
GPT-6.1 Sol 1,248
Grok 4.7 1,243
MiniMax-M3.1-Flash 1,243
GPT-5.6 Sol medium, xhigh 1,233
Grok 4.6 1,230
GLM-5.3 1,226
Claude Opus 4.8 1,218
Gemini 3.8 Flash 1,217
Gemini 3.7 Flash 1,216
Claude Sonnet 5 1,215
Grok 4.5 1,213
GPT-6 Sol 1,209
Ling-3.1-flash 1,198
GPT-6.1 Terra 1,193
Claude Opus 4.6 1,190
GLM-5.2 1,187
GPT-5.6 Sol medium 1,185
MiniMax-M3 1,180
Claude Sonnet 4.6 1,179
Muse Spark 1.2 1,178
Qwen3.7-Max 1,171
Gemini 3.6 Flash 1,169
Gemini 3.5 Flash 1,166
Muse Spark 1.1 1,165
Inkling 1,165
Muse Wonder max 1,162
GLM-5.1 1,154
Kimi K2.6 1,151
Kimi K2.7 Code 1,151
Claude Opus 4.5 1,149
GPT-6 Luna 1,146
GPT-5.6 Luna 1,141
Kimi K2.5 thinking 1,109
GLM-5 1,108
GPT-5.5 1,106
GLM-5V-Turbo 1,103
Grok 4.20 reasoning 1,086
GPT-5.6 Terra 1,083
Gemini 3.1 Pro 1,080
Solar Pro 4 1,076
Gemini 3 Flash 1,074
Gemini 3 Pro 1,073
Claude Sonnet 4.5 1,059
GPT-5.2 1,057
GLM-4.7 1,055
GLM-4.6 1,052
GPT-5-Codex 1,043
Grok 4.3 1,035
GPT-5.4 1,034
GPT-5.2 Codex 1,032
GPT-5.1-Codex 1,027
GPT-5.3 Codex 1,010
grok-code-fast-1 1,007
GPT-5.1 1,001
Grok 4.1 Fast reasoning 933
DeepSeek-V4-Pro 919
Mistral Large 908
Gemini 2.5 Pro Jun 2025 907
Mistral Medium 3.5 887

12 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.

Mistral AI Other labs OpenAI Anthropic Google DeepMind
Own leaderboard 12 leaders
Claude Opus 5.5 1,442

Source: Official DesignArena Agentic Overall 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 2024 2025 2026 2027 800 1,000 1,200 1,400 1,600 2024 2025 2026 2027

Source: Official DesignArena Agentic Overall 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