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

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

#ModelScore
330PaLM 2Closed. Closed weights: ranked by its LMArena score, with no size or downloads to show.887
329Llama-3.2-1B-InstructMeta Llama959
328Llama-2-7b-chat-hfMeta Llama973
327gemma-2b-itGoogle986
326Llama-2-70b-chat-hfMeta Llama995
325Llama-2-13b-chat-hfMeta Llama1001
324Mistral-7B-Instruct-v0.2Mistral AI1009
323gemma-1.1-2b-itGoogle1012
322GPT-3.5 Turbo (1106)Closed. Closed weights: ranked by its LMArena score, with no size or downloads to show.1012
321OLMo-7B-InstructAi21014
320Phi-3-mini-128k-instructMicrosoft1016
319Llama-3.2-3B-InstructMeta Llama1017
318Qwen1.5-4B-ChatQwen1024
317Phi-3-mini-4k-instructMicrosoft1034
316gemma-7b-itGoogle1036
315Mixtral-8x7B-Instruct-v0.1Mistral AI1055
314gemma-1.1-7b-itGoogle1061
313Phi-3-small-8k-instructMicrosoft1061
312granite-3.0-8b-instructIBM Granite1063
311granite-3.0-2b-instructIBM Granite1070
Show all 330 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.