Through an API

Open and closed models answer the same suite-standard 1.0 through their APIs, ranked together. 1 ranked model.

#ModelPass rate
1DeepSeek V4.1 Flash 1017Open weights100%

Model leaderboard

86 models in creative writing ranked by LMArena score · updated Nov 11, 2026

#ModelScore
86Llama-3.2-1B-InstructMeta Llama1033
85granite-3.0-2b-instructIBM Granite1046
84granite-3.0-8b-instructIBM Granite1071
83Llama-3.2-3B-InstructMeta Llama1094
82granite-3.1-2b-instructIBM Granite1116
81granite-3.1-8b-instructIBM Granite1129
80QwQ-32B-PreviewQwen1134
79Mixtral-8x22B-Instruct-v0.1Mistral AI1141
78Llama-3.1-8B-InstructMeta Llama1154
77Qwen2.5-Coder-32B-InstructQwen1174
76gpt-oss-20bOpenAI1202
75granite-4.2-3bIBM Granite1204
74granite-4.0-h-smallIBM Granite1210
73Qwen2.5-72B-InstructQwen1221
72Llama-3.1-70B-InstructMeta Llama1232
71granite-4.2-8bIBM Granite1235
70Llama-4-Scout-17B-16E-InstructMeta Llama1249
69Llama-3.3-70B-InstructMeta Llama1250
68granite-4.1-8bIBM Granite1251
67NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA1259
Show all 86 models

Score by size

Plot against
Best for its size

Scores: LMArena, CC BY 4.0. Measured on served versions, not local builds. A model with several served variants shows its best.