New Site Scores AI Models on Human IQ Scale

A new startup project called AI IQ is stirring up the tech community by applying a human-centric measurement—the IQ test—to artificial intelligence. The platform has assigned estimated intelligence quotients to over 50 of the world's most powerful language models, ranking them on a scale traditionally used for human cognitive ability. The results have already divided opinion. Some models scored surprisingly high, rivaling or even exceeding average human IQ scores, while others fell short. The project's creators argue that using a familiar metric like IQ makes AI capabilities more accessible and understandable to the general public. Instead of technical jargon like 'parameter count' or 'perplexity scores,' users can now see a simple number that suggests how 'smart' a model is. However, the methodology has drawn sharp criticism from AI researchers. Many argue that human IQ tests measure specific cognitive abilities like logical reasoning, spatial awareness, and verbal comprehension—traits that may not translate directly to AI performance. An AI that excels at mathematical reasoning might score high on an IQ test but fail miserably at tasks requiring common sense or emotional intelligence. Critics also point out that AI models are fundamentally different from human minds. They do not 'think' or 'understand' in the human sense; they process patterns in data. Reducing their capabilities to a single number could be misleading and could fuel unrealistic expectations about what AI can and cannot do. Proponents of AI IQ counter that any measurement is better than none. They argue that the project sparks important conversations about how we define intelligence in the age of machines. Is intelligence the ability to solve problems? To learn quickly? To adapt to new situations? The debate is far from settled. Regardless of where one stands, AI IQ has succeeded in one thing: making people think critically about how we evaluate artificial intelligence. As AI becomes more integrated into daily life, finding fair and accurate ways to measure its capabilities will only become more important.