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

141 models in chinese ranked by LMArena score · updated Nov 11, 2026

#ModelScore
141Llama-3.2-1B-InstructMeta Llama959
140Llama-2-7b-chat-hfMeta Llama973
139gemma-2b-itGoogle986
138Llama-2-70b-chat-hfMeta Llama995
137Llama-2-13b-chat-hfMeta Llama1001
136Mistral-7B-Instruct-v0.2Mistral AI1009
135gemma-1.1-2b-itGoogle1012
134OLMo-7B-InstructAi21014
133Phi-3-mini-128k-instructMicrosoft1016
132Llama-3.2-3B-InstructMeta Llama1017
131Qwen1.5-4B-ChatQwen1024
130Phi-3-mini-4k-instructMicrosoft1034
129gemma-7b-itGoogle1036
128Mixtral-8x7B-Instruct-v0.1Mistral AI1055
127gemma-1.1-7b-itGoogle1061
126Phi-3-small-8k-instructMicrosoft1061
125granite-3.0-8b-instructIBM Granite1063
124granite-3.0-2b-instructIBM Granite1070
123Meta-Llama-3-8B-InstructMeta Llama1076
122Qwen-14B-ChatQwen1077
Show all 141 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.