AI3Radar

AI3 LLM Index

Global LLM composite leaderboard — capability (layered geometric fusion) and value (capability per blended dollar), with versioned cohorts and data availability.

ai3-llm-index/v1.1 · cohort v1.1-20260815 · Data time Aug 20, 2026

#ModelCapabilityValueBlended priceData available
101mimo-v2-omni76.569%
102kimi-k2-thinking-turbo76.319%
103gpt-5.6-luna-xhigh76.119%
104mistral-medium-3.576.059%
105gpt-5-chat75.799%
106nvidia-nemotron-3-ultra-550b-a55b-nvfp475.539%
107amazon-nova-experimental-chat-26-02-1075.279%
108deepseek-v3-275.029%
109glm-4.674.769%
110deepseek-v3.2-exp-thinking74.59%
111claude-opus-4-20250514-thinking-16k74.249%
112qwen3-max-2025-09-2373.989%
113deepseek-v4-flash-high-preview73.799%
114qwen3-235b-a22b-instruct-250773.739%
115deepseek-v3.2-thinking73.479%
116deepseek-v3.2-exp73.219%
117deepseek-r1-052872.959%
118grok-4-fast-chat72.699%
119kimi-k2-0905-preview72.449%
120kimi-k2-0711-preview72.189%
121ernie-5.0-preview-102271.929%
122deepseek-v3.1-terminus-thinking71.669%
123minimax-m2-771.419%
124deepseek-v3-171.159%
125qwen3.5-122b-a10b70.899%
126deepseek-v3.1-thinking70.639%
127amazon-nova-experimental-chat-26-01-1070.379%
128mistral-large-370.12$0.759%
129qwen3-vl-235b-a22b-instruct69.869%
130seed-thinking-1.569.8236%
131deepseek-v3.1-terminus69.69%
132gpt-4.1-2025-04-1469.349%
133claude-opus-4-2025051469.089%
134hunyuan-hy3-preview68.839%
135claude-haiku-4-5-2025100168.5758.17$29%
136claude-opus-468.5636%
137glm-4-568.3190.83$1.19%
138grok-3-preview-02-2468.059%
139o3-mini67.9718%
140grok-4-070967.799%
141gemini-2.5-flash67.549%
142mistral-medium-250867.289%
143qwen3.5-27b67.029%
144grok-4-fast-reasoning66.769%
145gemini-2.5-flash-preview-09-202566.519%
146gemini66.339%
147qwen3-235b-a22b-no-thinking66.259%
148gpt-5.4-nano-high65.999%
149o1-2024-12-1765.739%
150qwen3-235b-a22b65.5318%