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

327 models in math ranked by LMArena score · updated Nov 11, 2026

#ModelScore
327gemma-2b-itGoogle1009
326OLMo-7B-InstructAi21018
325Qwen1.5-4B-ChatQwen1026
324Llama-2-7b-chat-hfMeta Llama1042
323gemma-1.1-2b-itGoogle1047
322PaLM 2Closed. Closed weights: ranked by its LMArena score, with no size or downloads to show.1049
321SOLAR-10.7B-Instruct-v1.0Upstage1064
320Llama-2-13b-chat-hfMeta Llama1065
319gemma-7b-itGoogle1066
318Qwen-14B-ChatQwen1068
317Qwen1.5-7B-ChatQwen1080
316Mistral-7B-Instruct-v0.2Mistral AI1085
315Llama-3.2-1B-InstructMeta Llama1086
314Phi-3-mini-128k-instructMicrosoft1089
313Llama-2-70b-chat-hfMeta Llama1091
312gemma-1.1-7b-itGoogle1107
311deepseek-llm-67b-chatDeepSeek1108
310granite-3.0-2b-instructIBM Granite1117
309Qwen1.5-14B-ChatQwen1125
308Llama-3.2-3B-InstructMeta Llama1126
Show all 327 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.