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 coding ranked by LMArena score · updated Nov 11, 2026

#ModelScore
86Llama-3.2-1B-InstructMeta Llama1070
85granite-3.0-2b-instructIBM Granite1090
84Llama-3.2-3B-InstructMeta Llama1098
83granite-3.0-8b-instructIBM Granite1112
82granite-3.1-2b-instructIBM Granite1149
81QwQ-32B-PreviewQwen1155
80Mixtral-8x22B-Instruct-v0.1Mistral AI1165
79granite-3.1-8b-instructIBM Granite1186
78Llama-3.1-8B-InstructMeta Llama1195
77granite-4.0-h-smallIBM Granite1249
76Llama-3.1-70B-InstructMeta Llama1260
75Llama-3.3-70B-InstructMeta Llama1268
74Qwen2.5-Coder-32B-InstructQwen1276
73Llama-4-Scout-17B-16E-InstructMeta Llama1286
72Llama-3.1-405B-InstructMeta Llama1291
71Qwen2.5-72B-InstructQwen1292
70Llama-4-Maverick-17B-128E-InstructMeta Llama1301
69gpt-oss-20bOpenAI1306
68granite-4.1-8bIBM Granite1313
67QwQ-32BQwen1333
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.