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

147 models in coding ranked by LMArena score · updated Nov 11, 2026

#ModelScore
147Qwen1.5-4B-ChatQwen999
146Llama-2-7b-chat-hfMeta Llama1002
145gemma-2b-itGoogle1009
144OLMo-7B-InstructAi21016
143gemma-1.1-2b-itGoogle1034
142Phi-3-mini-128k-instructMicrosoft1039
141gemma-7b-itGoogle1048
140Llama-2-13b-chat-hfMeta Llama1062
139SOLAR-10.7B-Instruct-v1.0Upstage1067
138Llama-3.2-1B-InstructMeta Llama1070
137Qwen-14B-ChatQwen1071
136Llama-2-70b-chat-hfMeta Llama1079
135Mistral-7B-Instruct-v0.2Mistral AI1082
134gemma-1.1-7b-itGoogle1084
133granite-3.0-2b-instructIBM Granite1090
132Phi-3-mini-4k-instructMicrosoft1093
131deepseek-llm-67b-chatDeepSeek1095
130Llama-3.2-3B-InstructMeta Llama1098
129Phi-3-small-8k-instructMicrosoft1101
128Qwen1.5-7B-ChatQwen1107
Show all 147 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.