Prose Variety

Subcategory of Creative Writing. 146 models scored.

Model Leaderboard

All models ranked by their Prose Variety subcategory score.

# Model Prose Variety Creative Writing Overall
1 Writer: Palmyra X5 83.01% 83.95% 78.11%
2 Qwen3 235B A22B Instruct 2507 81.37% 84.81% 78.07%
3 GPT-5.4 81.26% 90.94% 84.31%
4 GPT-5.4 (Reasoning, Low) 80.88% 90.51% 90.91%
5 GPT-5.4 (Reasoning) 79.74% 91.17% 93.85%
6 Cydonia 24B V4.1 78.35% 74.19% 72.68%
7 Claude Sonnet 4.5 78.05% 84.19% 87.54%
8 Claude Opus 4 77.58% 83.79% 87.22%
9 GPT-5.5 (Reasoning, Low) 77.34% 90.24% 92.51%
10 Z.AI GLM 5 77.33% 83.63% 89.60%
11 Mistral Small 4 (Reasoning) 77.22% 81.67% 79.48%
12 Claude Sonnet 4 76.88% 79.21% 87.64%
13 GPT-5.5 76.57% 90.39% 89.37%
14 Claude Opus 4.5 76.11% 81.71% 89.60%
15 GPT-5.4 Mini 75.74% 88.10% 80.45%
16 GPT-5.5 (Reasoning) 75.69% 90.26% 93.72%
17 Z.AI GLM 5.1 75.67% 84.05% 93.74%
18 Z.AI GLM 5 Turbo 75.38% 84.66% 93.29%
19 Mistral Small 4 75.37% 81.12% 75.23%
20 Claude Haiku 4.5 75.25% 78.96% 83.36%
21 Claude Opus 4.7 75.05% 84.74% 89.90%
22 Claude Opus 4.6 (Reasoning) 75.04% 84.55% 95.06%
23 Grok 4.5 (Reasoning, High) 74.99% 87.43% 94.12%
24 Claude Sonnet 5 (Reasoning, Low) 74.99% 81.43% 90.16%
25 GPT-5.4 Mini (Reasoning, Low) 74.97% 87.72% 83.57%
26 GPT-5.4 Mini (Reasoning) 74.97% 88.66% 89.82%
27 Ministral 3 14B 74.94% 79.11% 70.45%
28 GPT-5.6 Terra (Reasoning) 74.76% 90.72% 92.49%
29 Mistral Medium 3.1 74.30% 81.70% 76.08%
30 Grok 4.20 (Reasoning) 74.25% 86.25% 90.87%
31 Claude Opus 4.6 74.14% 83.59% 92.31%
32 Grok 4.20 74.09% 83.44% 81.21%
33 Claude Opus 4.8 (Reasoning, Low) 74.07% 85.86% 91.89%
34 Z.AI GLM 5.2 (Reasoning, High) 73.96% 83.92% 93.41%
35 GPT-5.6 Terra 73.87% 90.92% 88.23%
36 DeepSeek V3 (2025-03-24) 73.86% 82.34% 79.93%
37 GPT-5.6 Sol (Reasoning) 73.84% 91.03% 95.15%
38 GPT-5.1 73.76% 87.20% 90.73%
39 Hermes 3 70B 73.70% 77.41% 69.74%
40 MiniMax M2.5 73.63% 81.21% 86.71%
41 DeepSeek V4 Pro (Reasoning) 73.38% 82.99% 89.28%
42 Claude Sonnet 5 73.22% 82.58% 87.34%
43 Grok 4.5 (Reasoning, Low) 73.15% 87.45% 90.94%
44 DeepSeek V4 Pro 73.07% 83.70% 82.05%
45 GPT-4o, Aug. 6th (temp=1) 73.05% 75.50% 81.28%
46 Mistral Large 2 72.75% 81.86% 81.50%
47 Claude Opus 4.8 (Reasoning) 72.75% 85.25% 92.33%
48 Claude Opus 4.7 (Reasoning) 72.71% 84.73% 92.53%
49 Claude Sonnet 4.6 72.52% 83.31% 90.66%
50 Claude Sonnet 5 (Reasoning) 72.46% 81.31% 90.40%
51 GPT-5.6 Sol 72.39% 91.04% 90.76%
52 Mistral Large 3 72.38% 81.21% 84.29%
53 Aion 3.0 72.32% 84.64% 88.78%
54 GPT-5.6 Luna (Reasoning) 72.24% 90.10% 90.61%
55 MiniMax M2.7 72.16% 81.70% 86.23%
56 Claude Sonnet 4.6 (Reasoning) 71.90% 83.09% 93.64%
57 GPT-5.4 Nano 71.90% 80.50% 72.16%
58 GPT-5.6 Luna 71.83% 89.69% 85.06%
59 Qwen 3 32B 71.76% 81.30% 79.37%
60 Gemma 3 27B 71.72% 78.79% 75.70%
61 GPT-5.4 Nano (Reasoning, Low) 71.57% 80.93% 77.46%
62 Gemma 3 12B 71.52% 75.38% 76.07%
63 GPT-4o Mini (temp=1) 71.50% 74.37% 77.82%
64 Aion 3.0 Mini 71.45% 84.60% 83.69%
65 Cohere Command R+ (Aug. 2024) 71.05% 77.70% 67.04%
66 MoonshotAI: Kimi K2.5 70.84% 81.35% 90.86%
67 Ministral 3 8B 70.61% 77.26% 69.98%
68 DeepSeek V4 Flash (Reasoning) 70.46% 83.03% 88.06%
69 ByteDance Seed 1.6 Flash 70.46% 81.51% 70.92%
70 MiniMax M3 70.33% 84.57% 90.45%
71 Hermes 3 405B 69.94% 80.92% 80.80%
72 DeepSeek V4 Flash 69.82% 83.42% 82.02%
73 GPT-4.1 69.77% 81.24% 86.82%
74 Ministral 8B 69.62% 76.87% 63.77%
75 GPT-5.4 Nano (Reasoning) 69.53% 80.97% 80.02%
76 Z.AI GLM 4.5 68.98% 76.56% 84.95%
77 Gemma 3 4B 68.92% 72.10% 66.33%
78 Gemini 3.5 Flash (Reasoning, Minimal) 68.64% 78.55% 85.88%
79 Qwen 3.5 397B A17B 68.63% 86.93% 91.09%
80 Xiaomi MIMO v2.5 Pro 68.44% 81.08% 86.05%
81 MoonshotAI: Kimi K2.6 68.21% 85.47% 92.57%
82 GPT-4.1 Mini 68.20% 74.52% 81.40%
83 GPT-5.2 67.99% 80.36% 89.45%
84 Qwen3.6 Max Preview 67.96% 88.42% 93.72%
85 Z.AI GLM 4.5 Air 67.71% 74.61% 80.74%
86 Qwen 3.5 Plus (2026-02-15) 67.63% 77.07% 86.17%
87 Ministral 3B 67.17% 75.49% 59.25%
88 Qwen 3.6 Flash 67.13% 86.02% 89.31%
89 GPT-4.1 Nano 67.03% 71.81% 69.90%
90 DeepSeek-V2 Chat 66.87% 77.20% 84.09%
91 Gemini 2.5 Flash (Reasoning) 66.66% 76.30% 84.14%
92 Gemini 3.5 Flash (Reasoning) 66.61% 79.87% 93.35%
93 Gemini 2.5 Flash Lite 66.56% 75.05% 79.91%
94 Xiaomi MIMO v2.5 66.42% 79.16% 83.95%
95 DeepSeek V3 (2024-12-26) 66.31% 77.88% 82.62%
96 DeepSeek V3.2 66.11% 79.95% 82.22%
97 Gemini 2.5 Flash Lite (Reasoning) 65.69% 71.64% 83.10%
98 Llama 3.1 70B 65.53% 72.78% 77.41%
99 Gemini 2.5 Flash 65.37% 77.57% 80.61%
100 WizardLM 2 8x22b 65.28% 79.06% 71.45%
101 Ministral 3 3B 65.17% 75.45% 65.02%
102 Qwen 3.6 35B 65.03% 85.97% 87.66%
103 Aion 2.0 64.90% 80.24% 86.66%
104 Qwen 3.5 Plus (2026-04-20) 64.62% 85.18% 89.79%
105 o4 Mini 63.95% 82.04% 86.56%
106 Z.AI GLM 4.6 63.85% 78.86% 87.64%
107 o4 Mini High 63.78% 82.72% 88.78%
108 Gemini 3.1 Pro (Preview) 63.42% 85.44% 94.08%
109 Qwen 3.6 27B 63.42% 82.81% 88.33%
110 DeepSeek V3.1 62.92% 77.45% 82.35%
111 Gemini 2.5 Pro 62.91% 81.03% 88.44%
112 Grok 4.3 62.78% 84.51% 78.00%
113 GPT-5 62.58% 86.87% 91.48%
114 Gemma 4 31B 62.09% 75.59% 85.23%
115 Gemini 3 Flash (Preview, Reasoning) 61.94% 75.87% 89.93%
116 Qwen3.7 Max 61.91% 85.39% 94.55%
117 Z.AI GLM 4.7 Flash 61.69% 77.36% 82.21%
118 Z.AI GLM 4.7 61.49% 78.89% 87.67%
119 Gemini 3 Flash (Preview) 61.35% 75.04% 85.47%
120 GPT-5 Mini 60.91% 80.48% 91.31%
121 GPT-5 Nano 60.85% 67.04% 80.16%
122 Gemma 4 31B (Reasoning) 60.53% 78.13% 89.64%
123 Qwen 2.5 72B 60.17% 75.16% 73.17%
124 Mistral NeMO 59.98% 76.72% 63.80%
125 Gemini 3.1 Flash Lite (Preview) 59.92% 75.78% 85.41%
126 Gemma 4 26B (Reasoning) 59.82% 76.38% 89.02%
127 Gemma 4 26B 59.55% 75.17% 84.89%
128 GPT-4o Mini (temp=0) 59.43% 73.10% 76.86%
129 Gemini 3.1 Flash Lite (Reasoning) 59.35% 76.31% 85.91%
130 ByteDance Seed 2.0 Mini 59.35% 80.11% 85.69%
131 Grok 4.3 (Reasoning) 58.99% 85.11% 90.99%
132 GPT-4o, Aug. 6th (temp=0) 58.95% 73.65% 82.18%
133 Qwen 3.5 Flash 58.83% 83.81% 85.66%
134 Gemini 3.1 Flash Lite 58.53% 76.01% 85.09%
135 ByteDance Seed 2.0 Lite 58.28% 82.35% 84.27%
136 Qwen 3.5 35B 58.12% 83.51% 87.01%
137 Nemotron 3 Super 57.81% 69.75% 81.69%
138 ByteDance Seed 1.6 57.70% 78.43% 89.59%
139 Arcee AI: Trinity Mini 56.69% 74.01% 67.68%
140 Qwen 3.5 122B 56.46% 83.02% 90.32%
141 Qwen 3.5 27B 54.98% 82.54% 90.05%
142 Qwen 3.5 9B 53.95% 84.35% 84.05%
143 Nemotron 3 Nano 52.17% 65.87% 74.50%
144 GPT-OSS 120B 50.96% 67.85% 84.81%
145 Inception Mercury 2 50.37% 68.31% 81.99%
146 Mistral Small 3.2 24B 47.89% 71.87% 77.36%