Benchmark Almanac

Creative Writing v3

Kimi K3.1 Pro 83.2 leads since Oct 2026 Dropped

Models write responses to 32 creative writing prompts, three times each. A judge model scores them with a rubric and pairwise Elo comparisons.

Version 3 replaced version 2 after the judge could no longer separate top models.

Made by EQ-Bench Released Mar 2025 Creative writing All tasks public Last in a launch post Nov 2025 , dropped before it saturated Last model added Oct 2026 Scores refreshed 2 Nov 2026
Level
University Undergraduates and entry-level professionals
Scored
Graded by a model A language model scores each response, usually against a rubric.
Answers
Open-ended Free-form text, documents, designs or media judged as a whole.
EQ-bench/creative-writing-bench 151 stars Leaderboard eqbench.com
144
Entries on its own leaderboard, 120 models we track
8
Launch posts list it, from 5 labs
11
System cards report it, from 3 labs

Kimi K3.1 Pro leads with 83.2

Best score per entry from the benchmark's own leaderboard, read on 2 Nov 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.

144 entries. Scroll sideways for the rest.

  1. 83.2 Kimi K3.1 Pro
  2. 82.2 GLM-5.3
  3. 82.2 ox-alpha
  4. 81.9 Claude Fable 5
  5. 81.8 Kimi K3
  6. 81.3 Claude Fable 5.1
  7. 81.3 Claude Opus 5
  8. 81.1 Muse Wonder
  9. 80.8 GPT-5.6 Sol
  10. 80.7 Claude Opus 4.7
  11. 80.6 aion-3.5
  12. 80.3 GPT-6 Sol
  13. 79.9 Claude Opus 5.5
  14. 79.8 Muse Spark 1.1
  15. 79.8 Grok 4.7
  16. 79.7 Claude Opus 4.8
  17. 79.7 GPT-6 Astra
  18. 79.5 GPT-5.6 Terra
  19. 79.5 GPT-5.5
  20. 78.8 GPT-5.4
  21. 78.5 GPT-5.6 Luna
  22. 78.3 Claude Opus 4.6
  23. 78.2 Muse Spark 1.3
  24. 78.0 Gemini 3.8 Flash
  25. 78.0 Muse Spark 1.2
  26. 77.8 Gemini 4 Argon
  27. 77.7 Claude Sonnet 4.6
  28. 77.3 Altworld/Hemmingway-1
  29. 76.6 GLM-5.2
  30. 76.5 Qwen3.8-2.4T-A95B
  31. 76.3 Claude Sonnet 5
  32. 76.1 Gemini 3.7 Flash
  33. 75.0 Kimi K2
  34. 75.0 Claude Opus 4.5
  35. 75.0 DeepSeek-V4-Pro
  36. 74.7 Kimi K2.6
  37. 74.3 Gemini 3 Pro
  38. 74.0 GPT-5.2
  39. 73.5 nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4
  40. 73.3 Kimi K2 Thinking
  41. 73.3 openrouter/horizon-beta
  42. 73.3 o3
  43. 72.7 GPT-5
  44. 72.5 GPT-5.4 Mini
  45. 72.5 Gemini 3.6 Flash
  46. 72.3 Gemini 2.5 Pro
  47. 72.3 Gemini 3.5 Flash-Lite
  48. 72.2 Claude Sonnet 4.5
  49. 72.2 GLM-5.1
  50. 71.9 openrouter/pony-alpha
  51. 71.7 DeepSeek-V3.2
  52. 71.4 MiMo-V2.5-Pro
  53. 71.2 Gemini 3.1 Pro
  54. 71.0 GPT-5.3 Instant
  55. 70.9 Inkling
  56. 70.7 hunter-alpha
  57. 70.6 GLM-5
  58. 70.5 Grok 4.5
  59. 70.4 Kimi K2.5
  60. 70.3 DeepSeek V4.1 Flash
  61. 70.2 GLM-4.7
  62. 70.0 Claude Opus 4
  63. 69.9 DeepSeek-V3.1
  64. 69.8 DeepSeek-V4-Flash
  65. 69.3 GLM-4.6
  66. 68.8 openrouter/horizon-alpha
  67. 68.8 GPT-4o
  68. 68.6 meta-models/Muse-Glimmer-30B
  69. 68.3 DeepSeek-R1
  70. 68.2 space-bunny-alpha
  71. 68.0 Gemma 4 31B IT
  72. 67.5 DeepSeek V4 Flash 0731
  73. 66.7 Hermes 4 405B
  74. 66.3 Qwen3.5 397B-A17B
  75. 66.3 Mistral Medium 3.1
  76. 65.8 Gemma 4 26B A4B
  77. 65.8 DeepSeek-V3
  78. 65.8 GPT-4.1
  79. 65.7 Grok 4.20
  80. 65.4 Claude Sonnet 4
  81. 64.8 Qwen 3.8 27B
  82. 64.7 Claude 3.7 Sonnet
  83. 64.5 optimus-alpha
  84. 64.2 Claude 3.5 Sonnet
  85. 63.8 Qwen3-235B-A22B-Thinking
  86. 63.0 GLM-4.5
  87. 63.0 GPT-4o
  88. 62.3 Mistral Large 3
  89. 62.3 quasar-alpha
  90. 62.2 Grok 3
  91. 61.9 Gemma 4 12B
  92. 61.8 AlexWortega/Qwen3.8-27B-antislop
  93. 61.4 Grok 4.1 Fast
  94. 61.2 openrouter/sherlock-dash-alpha
  95. 60.9 mistral-small-creative
  96. 60.6 GPT-4.5
  97. 60.2 MiniMax-M2.5
  98. 60.0 DeepSeek-V3.2-Speciale
  99. 60.0 GPT-5 mini
  100. 58.9 Inkling-Small
  101. 58.7 Gemma 3 27B
  102. 58.7 Gemini 2.5 Flash
  103. 58.0 CohereForAI/c4ai-command-a-03-2025
  104. 58.0 Gemma 4 E4B
  105. 58.0 Nemotron 3.5 Lightning 30B-A3B
  106. 57.2 Mistral Small 3.2
  107. 57.0 Gemini 2.0 Flash
  108. 57.0 Nemotron 3.5 Bolt E4B
  109. 57.0 Gemma 3 12B
  110. 56.8 allura-org/Gemma-3-Glitter-12B
  111. 56.5 QwQ-32B
  112. 55.5 GPT-4.1 mini
  113. 55.2 glm-4.7-flash
  114. 54.8 Reka Flash 3
  115. 54.1 Gemma 3 4B
  116. 54.1 Claude 3.5 Haiku
  117. 53.5 sam-paech/Darkest-muse-v1
  118. 53.0 THUDM/GLM-4-32B-0414
  119. 52.5 ifable/gemma-2-Ifable-9B
  120. 51.9 GPT-4o mini
  121. 51.9 Mistral Large 2
  122. 51.7 Pixtral Large
  123. 51.7 Llama 4 Maverick
  124. 51.5 Llama 3.1-405B
  125. 50.9 GPT-4.1 nano
  126. 50.8 GPT-4
  127. 50.8 Llama 3.1-70B
  128. 49.5 Gemma 2 9B
  129. 48.3 Claude 3 Haiku
  130. 48.0 Mistral NeMo
  131. 47.9 gpt-oss-120b
  132. 47.5 ToastyPigeon/Gemma-3-Starshine-12B
  133. 47.2 Llama 3.1-8B
  134. 46.7 liquid/lfm-7b
  135. 46.6 Llama 4 Scout
  136. 45.9 GPT-5 nano
  137. 45.0 openrouter/cypher-alpha
  138. 44.5 Nanbeige/Nanbeige4-3B-Thinking-2511
  139. 42.8 Llama 3.2 3B
  140. 42.5 GPT-3.5 Turbo
  141. 42.4 Mistral Small 3
  142. 42.3 Mistral Small 3.1
  143. 40.5 gpt-oss-20b
  144. 31.9 Llama 3.2 1B
  1. 1 Kimi K3.1 Pro 83.2
  2. 2 GLM-5.3 82.2
  3. 3 ox-alpha 82.2
  4. 4 Claude Fable 5 81.9
  5. 5 Kimi K3 81.8
  6. 6 Claude Fable 5.1 81.3
  7. 7 Claude Opus 5 81.3
  8. 8 Muse Wonder 81.1
  9. 9 GPT-5.6 Sol 80.8
  10. 10 Claude Opus 4.7 80.7
  11. 11 aion-3.5 80.6
  12. 12 GPT-6 Sol 80.3
  13. 13 Claude Opus 5.5 79.9
  14. 14 Muse Spark 1.1 79.8
  15. 15 Grok 4.7 79.8
  16. 16 Claude Opus 4.8 79.7
  17. 17 GPT-6 Astra 79.7
  18. 18 GPT-5.6 Terra 79.5
  19. 19 GPT-5.5 79.5
  20. 20 GPT-5.4 78.8
  21. 21 GPT-5.6 Luna 78.5
  22. 22 Claude Opus 4.6 78.3
  23. 23 Muse Spark 1.3 78.2
  24. 24 Gemini 3.8 Flash 78.0
  25. 25 Muse Spark 1.2 78.0
  26. 26 Gemini 4 Argon 77.8
  27. 27 Claude Sonnet 4.6 77.7
  28. 28 Altworld/Hemmingway-1 77.3
  29. 29 GLM-5.2 76.6
  30. 30 Qwen3.8-2.4T-A95B 76.5
  31. 31 Claude Sonnet 5 76.3
  32. 32 Gemini 3.7 Flash 76.1
  33. 33 Kimi K2 Jul 2025 75.0
  34. 34 Claude Opus 4.5 75.0
  35. 35 DeepSeek-V4-Pro 75.0
  36. 36 Kimi K2.6 74.7
  37. 37 Gemini 3 Pro 74.3
  38. 38 GPT-5.2 74.0
  39. 39 nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 73.5
  40. 40 Kimi K2 Thinking 73.3
  41. 41 openrouter/horizon-beta 73.3
  42. 42 o3 73.3
  43. 43 GPT-5 72.7
  44. 44 GPT-5.4 Mini 72.5
  45. 45 Gemini 3.6 Flash 72.5
  46. 46 Gemini 2.5 Pro Jun 2025 72.3
  47. 47 Gemini 3.5 Flash-Lite 72.3
  48. 48 Claude Sonnet 4.5 72.2
  49. 49 GLM-5.1 72.2
  50. 50 openrouter/pony-alpha 71.9
  51. 51 DeepSeek-V3.2 71.7
  52. 52 MiMo-V2.5-Pro 71.4
  53. 53 Gemini 3.1 Pro 71.2
  54. 54 GPT-5.3 Instant 71.0
  55. 55 Inkling 70.9
  56. 56 hunter-alpha 70.7
  57. 57 GLM-5 70.6
  58. 58 Grok 4.5 70.5
  59. 59 Kimi K2.5 70.4
  60. 60 DeepSeek V4.1 Flash 70.3
  61. 61 GLM-4.7 70.2
  62. 62 Claude Opus 4 70.0
  63. 63 DeepSeek-V3.1 69.9
  64. 64 DeepSeek-V4-Flash 69.8
  65. 65 GLM-4.6 69.3
  66. 66 openrouter/horizon-alpha 68.8
  67. 67 GPT-4o Mar 2025 68.8
  68. 68 meta-models/Muse-Glimmer-30B 68.6
  69. 69 DeepSeek-R1 May 2025 68.3
  70. 70 space-bunny-alpha 68.2
  71. 71 Gemma 4 31B IT 68.0
  72. 72 DeepSeek V4 Flash 0731 67.5
  73. 73 Hermes 4 405B 66.7
  74. 74 Qwen3.5 397B-A17B 66.3
  75. 75 Mistral Medium 3.1 66.3
  76. 76 Gemma 4 26B A4B 65.8
  77. 77 DeepSeek-V3 Mar 2025 65.8
  78. 78 GPT-4.1 65.8
  79. 79 Grok 4.20 65.7
  80. 80 Claude Sonnet 4 65.4
  81. 81 Qwen 3.8 27B 64.8
  82. 82 Claude 3.7 Sonnet 64.7
  83. 83 optimus-alpha 64.5
  84. 84 Claude 3.5 Sonnet 64.2
  85. 85 Qwen3-235B-A22B-Thinking Jul 2025 63.8
  86. 86 GLM-4.5 63.0
  87. 87 GPT-4o Jan 2025 63.0
  88. 88 Mistral Large 3 62.3
  89. 89 quasar-alpha 62.3
  90. 90 Grok 3 62.2
  91. 91 Gemma 4 12B 61.9
  92. 92 AlexWortega/Qwen3.8-27B-antislop 61.8
  93. 93 Grok 4.1 Fast 61.4
  94. 94 openrouter/sherlock-dash-alpha 61.2
  95. 95 mistral-small-creative 60.9
  96. 96 GPT-4.5 60.6
  97. 97 MiniMax-M2.5 60.2
  98. 98 DeepSeek-V3.2-Speciale 60.0
  99. 99 GPT-5 mini 60.0
  100. 100 Inkling-Small 58.9
  101. 101 Gemma 3 27B 58.7
  102. 102 Gemini 2.5 Flash Sep 2025 58.7
  103. 103 CohereForAI/c4ai-command-a-03-2025 58.0
  104. 104 Gemma 4 E4B 58.0
  105. 105 Nemotron 3.5 Lightning 30B-A3B 58.0
  106. 106 Mistral Small 3.2 57.2
  107. 107 Gemini 2.0 Flash Feb 2025 57.0
  108. 108 Nemotron 3.5 Bolt E4B 57.0
  109. 109 Gemma 3 12B 57.0
  110. 110 allura-org/Gemma-3-Glitter-12B 56.8
  111. 111 QwQ-32B 56.5
  112. 112 GPT-4.1 mini 55.5
  113. 113 glm-4.7-flash 55.2
  114. 114 Reka Flash 3 54.8
  115. 115 Gemma 3 4B 54.1
  116. 116 Claude 3.5 Haiku 54.1
  117. 117 sam-paech/Darkest-muse-v1 53.5
  118. 118 THUDM/GLM-4-32B-0414 53.0
  119. 119 ifable/gemma-2-Ifable-9B 52.5
  120. 120 GPT-4o mini 51.9
  121. 121 Mistral Large 2 Nov 2024 51.9
  122. 122 Pixtral Large 51.7
  123. 123 Llama 4 Maverick 51.7
  124. 124 Llama 3.1-405B 51.5
  125. 125 GPT-4.1 nano 50.9
  126. 126 GPT-4 Mar 2023 50.8
  127. 127 Llama 3.1-70B 50.8
  128. 128 Gemma 2 9B 49.5
  129. 129 Claude 3 Haiku 48.3
  130. 130 Mistral NeMo 48.0
  131. 131 gpt-oss-120b 47.9
  132. 132 ToastyPigeon/Gemma-3-Starshine-12B 47.5
  133. 133 Llama 3.1-8B 47.2
  134. 134 liquid/lfm-7b 46.7
  135. 135 Llama 4 Scout 46.6
  136. 136 GPT-5 nano 45.9
  137. 137 openrouter/cypher-alpha 45.0
  138. 138 Nanbeige/Nanbeige4-3B-Thinking-2511 44.5
  139. 139 Llama 3.2 3B 42.8
  140. 140 GPT-3.5 Turbo Jun 2023 42.5
  141. 141 Mistral Small 3 42.4
  142. 142 Mistral Small 3.1 42.3
  143. 143 gpt-oss-20b 40.5
  144. 144 Llama 3.2 1B 31.9
#EntryScoreSettingDateSource
1 Kimi K3.1 Pro 83.2 Oct 2026 Official leaderboard
2 GLM-5.3 82.2 Aug 2026 Official leaderboard
3 ox-alpha 82.2 Official leaderboard
4 Claude Fable 5 81.9 Jun 2026 Official leaderboard
5 Kimi K3 81.8 Jul 2026 Official leaderboard
6 Claude Fable 5.1 81.3 Sep 2026 Official leaderboard
7 Claude Opus 5 81.3 Jul 2026 Official leaderboard
8 Muse Wonder 81.1 Oct 2026 Official leaderboard
9 GPT-5.6 Sol 80.8 Jul 2026 Official leaderboard
10 Claude Opus 4.7 80.7 Apr 2026 Official leaderboard
11 aion-3.5 80.6 Official leaderboard
12 GPT-6 Sol 80.3 Sep 2026 Official leaderboard
13 Claude Opus 5.5 79.9 Sep 2026 Official leaderboard
14 Muse Spark 1.1 79.8 Jul 2026 Official leaderboard
15 Grok 4.7 79.8 Sep 2026 Official leaderboard
16 Claude Opus 4.8 79.7 May 2026 Official leaderboard
17 GPT-6 Astra 79.7 Sep 2026 Official leaderboard
18 GPT-5.6 Terra 79.5 Jul 2026 Official leaderboard
19 GPT-5.5 79.5 Apr 2026 Official leaderboard
20 GPT-5.4 78.8 Mar 2026 Official leaderboard
21 GPT-5.6 Luna 78.5 Jul 2026 Official leaderboard
22 Claude Opus 4.6 78.3 Feb 2026 Official leaderboard
23 Muse Spark 1.3 78.2 Sep 2026 Official leaderboard
24 Gemini 3.8 Flash 78.0 Sep 2026 Official leaderboard
25 Muse Spark 1.2 78.0 Aug 2026 Official leaderboard
26 Gemini 4 Argon 77.8 Sep 2026 Official leaderboard
27 Claude Sonnet 4.6 77.7 Feb 2026 Official leaderboard
28 Altworld/Hemmingway-1 77.3 Official leaderboard
29 GLM-5.2 76.6 Jun 2026 Official leaderboard
30 Qwen3.8-2.4T-A95B 76.5 Aug 2026 Official leaderboard
31 Claude Sonnet 5 76.3 Jun 2026 Official leaderboard
32 Gemini 3.7 Flash 76.1 Aug 2026 Official leaderboard
33 Kimi K2 Jul 2025 75.0 Jul 2025 Official leaderboard
34 Claude Opus 4.5 75.0 Nov 2025 Official leaderboard
35 DeepSeek-V4-Pro 75.0 Apr 2026 Official leaderboard
36 Kimi K2.6 74.7 Apr 2026 Official leaderboard
37 Gemini 3 Pro 74.3 Nov 2025 Official leaderboard
38 GPT-5.2 74.0 Dec 2025 Official leaderboard
39 nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 73.5 Official leaderboard
40 Kimi K2 Thinking 73.3 Nov 2025 Official leaderboard
41 openrouter/horizon-beta 73.3 Official leaderboard
42 o3 73.3 Apr 2025 Official leaderboard
43 GPT-5 72.7 Aug 2025 Official leaderboard
44 GPT-5.4 Mini 72.5 Mar 2026 Official leaderboard
45 Gemini 3.6 Flash 72.5 Jul 2026 Official leaderboard
46 Gemini 2.5 Pro Jun 2025 72.3 Jun 2025 Official leaderboard
47 Gemini 3.5 Flash-Lite 72.3 Jul 2026 Official leaderboard
48 Claude Sonnet 4.5 72.2 Sep 2025 Official leaderboard
49 GLM-5.1 72.2 Apr 2026 Official leaderboard
50 openrouter/pony-alpha 71.9 Official leaderboard
51 DeepSeek-V3.2 71.7 Dec 2025 Official leaderboard
52 MiMo-V2.5-Pro 71.4 Apr 2026 Official leaderboard
53 Gemini 3.1 Pro 71.2 Feb 2026 Official leaderboard
54 GPT-5.3 Instant 71.0 Mar 2026 Official leaderboard
55 Inkling 70.9 Jul 2026 Official leaderboard
56 hunter-alpha 70.7 Official leaderboard
57 GLM-5 70.6 Feb 2026 Official leaderboard
58 Grok 4.5 70.5 Jul 2026 Official leaderboard
59 Kimi K2.5 70.4 Jan 2026 Official leaderboard
60 DeepSeek V4.1 Flash 70.3 Sep 2026 Official leaderboard
61 GLM-4.7 70.2 Dec 2025 Official leaderboard
62 Claude Opus 4 70.0 May 2025 Official leaderboard
63 DeepSeek-V3.1 69.9 Aug 2025 Official leaderboard
64 DeepSeek-V4-Flash 69.8 Apr 2026 Official leaderboard
65 GLM-4.6 69.3 Sep 2025 Official leaderboard
66 openrouter/horizon-alpha 68.8 Official leaderboard
67 GPT-4o Mar 2025 68.8 Mar 2025 Official leaderboard
68 meta-models/Muse-Glimmer-30B 68.6 Official leaderboard
69 DeepSeek-R1 May 2025 68.3 May 2025 Official leaderboard
70 space-bunny-alpha 68.2 Official leaderboard
71 Gemma 4 31B IT 68.0 Apr 2026 Official leaderboard
72 DeepSeek V4 Flash 0731 67.5 Jul 2026 Official leaderboard
73 Hermes 4 405B 66.7 Aug 2025 Official leaderboard
74 Qwen3.5 397B-A17B 66.3 Feb 2026 Official leaderboard
75 Mistral Medium 3.1 66.3 Aug 2025 Official leaderboard
76 Gemma 4 26B A4B 65.8 Apr 2026 Official leaderboard
77 DeepSeek-V3 Mar 2025 65.8 Mar 2025 Official leaderboard
78 GPT-4.1 65.8 Apr 2025 Official leaderboard
79 Grok 4.20 65.7 Feb 2026 Official leaderboard
80 Claude Sonnet 4 65.4 May 2025 Official leaderboard
81 Qwen 3.8 27B 64.8 Aug 2026 Official leaderboard
82 Claude 3.7 Sonnet 64.7 Feb 2025 Official leaderboard
83 optimus-alpha 64.5 Official leaderboard
84 Claude 3.5 Sonnet 64.2 Jun 2024 Official leaderboard
85 Qwen3-235B-A22B-Thinking Jul 2025 63.8 Jul 2025 Official leaderboard
86 GLM-4.5 63.0 Aug 2025 Official leaderboard
87 GPT-4o Jan 2025 63.0 Jan 2025 Official leaderboard
88 Mistral Large 3 62.3 Dec 2025 Official leaderboard
89 quasar-alpha 62.3 Official leaderboard
90 Grok 3 62.2 Apr 2025 Official leaderboard
91 Gemma 4 12B 61.9 Jun 2026 Official leaderboard
92 AlexWortega/Qwen3.8-27B-antislop 61.8 Official leaderboard
93 Grok 4.1 Fast 61.4 Nov 2025 Official leaderboard
94 openrouter/sherlock-dash-alpha 61.2 Official leaderboard
95 mistral-small-creative 60.9 Official leaderboard
96 GPT-4.5 60.6 Feb 2025 Official leaderboard
97 MiniMax-M2.5 60.2 Feb 2026 Official leaderboard
98 DeepSeek-V3.2-Speciale 60.0 Dec 2025 Official leaderboard
99 GPT-5 mini 60.0 Aug 2025 Official leaderboard
100 Inkling-Small 58.9 Jul 2026 Official leaderboard
101 Gemma 3 27B 58.7 Mar 2025 Official leaderboard
102 Gemini 2.5 Flash Sep 2025 58.7 Sep 2025 Official leaderboard
103 CohereForAI/c4ai-command-a-03-2025 58.0 Official leaderboard
104 Gemma 4 E4B 58.0 Mar 2026 Official leaderboard
105 Nemotron 3.5 Lightning 30B-A3B 58.0 Aug 2026 Official leaderboard
106 Mistral Small 3.2 57.2 Jun 2025 Official leaderboard
107 Gemini 2.0 Flash Feb 2025 57.0 Feb 2025 Official leaderboard
108 Nemotron 3.5 Bolt E4B 57.0 Oct 2026 Official leaderboard
109 Gemma 3 12B 57.0 Mar 2025 Official leaderboard
110 allura-org/Gemma-3-Glitter-12B 56.8 Official leaderboard
111 QwQ-32B 56.5 Mar 2025 Official leaderboard
112 GPT-4.1 mini 55.5 Apr 2025 Official leaderboard
113 glm-4.7-flash 55.2 Jan 2026 Official leaderboard
114 Reka Flash 3 54.8 Mar 2025 Official leaderboard
115 Gemma 3 4B 54.1 Mar 2025 Official leaderboard
116 Claude 3.5 Haiku 54.1 Oct 2024 Official leaderboard
117 sam-paech/Darkest-muse-v1 53.5 Official leaderboard
118 THUDM/GLM-4-32B-0414 53.0 Official leaderboard
119 ifable/gemma-2-Ifable-9B 52.5 Official leaderboard
120 GPT-4o mini 51.9 Jul 2024 Official leaderboard
121 Mistral Large 2 Nov 2024 51.9 Nov 2024 Official leaderboard
122 Pixtral Large 51.7 Nov 2024 Official leaderboard
123 Llama 4 Maverick 51.7 Apr 2025 Official leaderboard
124 Llama 3.1-405B 51.5 Jul 2024 Official leaderboard
125 GPT-4.1 nano 50.9 Apr 2025 Official leaderboard
126 GPT-4 Mar 2023 50.8 Mar 2023 Official leaderboard
127 Llama 3.1-70B 50.8 Jul 2024 Official leaderboard
128 Gemma 2 9B 49.5 Jun 2024 Official leaderboard
129 Claude 3 Haiku 48.3 Mar 2024 Official leaderboard
130 Mistral NeMo 48.0 Jul 2024 Official leaderboard
131 gpt-oss-120b 47.9 Aug 2025 Official leaderboard
132 ToastyPigeon/Gemma-3-Starshine-12B 47.5 Official leaderboard
133 Llama 3.1-8B 47.2 Jul 2024 Official leaderboard
134 liquid/lfm-7b 46.7 Official leaderboard
135 Llama 4 Scout 46.6 Apr 2025 Official leaderboard
136 GPT-5 nano 45.9 Aug 2025 Official leaderboard
137 openrouter/cypher-alpha 45.0 Official leaderboard
138 Nanbeige/Nanbeige4-3B-Thinking-2511 44.5 Official leaderboard
139 Llama 3.2 3B 42.8 Sep 2024 Official leaderboard
140 GPT-3.5 Turbo Jun 2023 42.5 Jun 2023 Official leaderboard
141 Mistral Small 3 42.4 Jan 2025 Official leaderboard
142 Mistral Small 3.1 42.3 Mar 2025 Official leaderboard
143 gpt-oss-20b 40.5 Aug 2025 Official leaderboard
144 Llama 3.2 1B 31.9 Sep 2024 Official leaderboard

Labs report scores for 12 models in their cards. We plot every source in the charts below.

11 models have held number one since its release in Mar 2025

Lab claims count from the day the card came out, independent runs from the model’s release. We draw one row each for the benchmark’s own leaderboard and what labs claim in their cards.

Anthropic DeepSeek OpenAI Moonshot AI Z.ai Qwen
Own leaderboard 11 leaders
Kimi K3.1 Pro 83.2
Lab claims Not tested since Nov 2025
Qwen3-235B-A22B-Thinking 86.1

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
20.0 40.0 60.0 80.0 100.0 2023 2024 2025 2026 2027 20.0 40.0 60.0 80.0 100.0 2023 2024 2025 2026 2027

No launch post has listed it since Nov 2025

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.

Qwen
Nous Research
Liquid AI
Ant Group
xAI
DateLaunch postScoreAs the post writes it
17 Nov 2025 Grok 4.1 1721.9 Creative Writing v3 (Elo (Normalized))
10 Oct 2025 Ring-1T 85.4 Creative Writing v3 (score)
7 Oct 2025 LFM2-8B-A1B 44.2 Creative Writing v3
23 Sep 2025 Qwen3-VL-235B-A22B 85.7 Creative Writing v3
22 Sep 2025 Qwen3-Omni-30B-A3B 80.6 Creative Writing v3
26 Aug 2025 Hermes 4 405B 79.8 CreativeWriting3 (reasoning)
25 Jul 2025 Qwen3-235B-A22B-Thinking 86.1 Creative Writing v3
29 Apr 2025 Qwen3-235B-A22B 84.6 Creative Writing v3

11 system cards report it, from 3 labs

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

DateDocumentScoreAs the document writes itSection
26 Nov 2025 Qwen3-VL-235B-A22B technical report 85.7 Creative Writing v3 (Table 6) Capabilities
21 Oct 2025 Ring-1T technical report 85.4 Creative Writing v3 (Table 1) Capabilities
22 Sep 2025 Qwen3-Omni-30B-A3B technical report 80.6 Creative Writing v3, Instruct (Table 4). Thinking 82.5 (Table 5) Capabilities
25 Aug 2025 Hermes 4 405B technical report 79.3 CreativeWriting3, reasoning mode (Table 3). Non-reasoning 50.6. v1 printed 79.8 Capabilities
14 May 2025 Qwen3-8B technical report 75 Creative Writing v3, thinking mode (Table 17). Non-thinking 64.5 (Table 18) Capabilities
14 May 2025 Qwen3-4B technical report 61.1 Creative Writing v3, thinking mode (Table 17). Non-thinking 53.6 (Table 18) Capabilities
14 May 2025 Qwen3-32B technical report 81 Creative Writing v3, thinking mode (Table 13). Non-thinking 78.3 (Table 14) Capabilities
14 May 2025 Qwen3-30B-A3B technical report 79.1 Creative Writing v3, thinking mode (Table 15). Non-thinking 68.1 (Table 16) Capabilities
14 May 2025 Qwen3-235B-A22B technical report 84.6 Creative Writing v3, thinking mode (Table 11). Non-thinking 80.4 (Table 12) Capabilities
14 May 2025 Qwen3-14B technical report 80.3 Creative Writing v3, thinking mode (Table 15). Non-thinking 73.1 (Table 16) Capabilities
14 May 2025 Qwen3-1.7B technical report 48 Creative Writing v3, thinking mode (Table 19). Non-thinking 43.6 (Table 20) Capabilities

More in Writing and taste

All 24 in Writing and taste
  • Design Bench (Moonshot) In use Models make graphic designs from direct prompts or few-shot reference images. An automated grader rates the results as Elo scores. Kimi K3.1 Pro 957
  • WritingBench Fading Models complete 1,000 real-world writing requests across 6 domains and 100 subdomains. A judge model scores each reply on five query-specific criteria. GPT-5 83.9
  • Creative Writing (MiniMax in-house) Only in system cards Models write creative pieces for real Hailuo AI user requests. Automatic assessments score the writing. MiniMax-Text-01 81.3%
  • DeepSeek Chinese functional writing evaluation Only in system cards Models answer 3,170 everyday Chinese writing prompts such as reports, emails and speeches. Pairwise judging against a baseline model gives a win rate. DeepSeek-V4-Pro 62.6%
  • Slide generation human evaluation (Z.ai) Only in system cards People compare HTML slide decks from two models on content, layout and visual quality. Metric is overall win rate. GLM-5 67.5%
  • Task-dependent output homogenization evaluation Only in system cards Models answer prompts several times. A judge scores functional diversity where it is wanted and consistency where one answer is right. Muse Spark 0.21