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

339 models in hard prompts ranked by LMArena score · updated Nov 11, 2026

#ModelScore
339Qwen1.5-4B-ChatQwen976
338gemma-2b-itGoogle989
337OLMo-7B-InstructAi2993
336PaLM 2Closed. Closed weights: ranked by its LMArena score, with no size or downloads to show.1005
335gemma-1.1-2b-itGoogle1005
334Llama-2-7b-chat-hfMeta Llama1009
333Qwen-14B-ChatQwen1027
332Phi-3-mini-128k-instructMicrosoft1028
331gemma-7b-itGoogle1042
330Llama-3.2-1B-InstructMeta Llama1044
329Llama-2-13b-chat-hfMeta Llama1051
328SOLAR-10.7B-Instruct-v1.0Upstage1062
327Qwen1.5-7B-ChatQwen1065
326Mistral-7B-Instruct-v0.2Mistral AI1067
325deepseek-llm-67b-chatDeepSeek1070
324gemma-1.1-7b-itGoogle1071
323Llama-2-70b-chat-hfMeta Llama1073
322granite-3.0-2b-instructIBM Granite1073
321Phi-3-mini-4k-instructMicrosoft1087
320granite-3.0-8b-instructIBM Granite1092
Show all 339 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.