AI vs. IQ: Just One Model Passed the Culture Fair Test

ChatGPT IQ Test Results

The results? Only one out of three models successfully completed the test within the time limit. However, GPT-o1 exceeded the time limit, resulting in a technical fail. The only model to both complete the test on time and pass was ChatGPT-4o.

Here's how the models performed on our official IQ test:

Both GPT-o1 and GPT-4o achieved a score of 82, placing them in the "Dullness" IQ score range. While the term may sound a bit striking, it simply indicates slightly below-average performance compared to the human average IQ of 100.

Interestingly, performance patterns revealed that ChatGPT-4o handled the more difficult questions toward the end better, while GPT-o1 excelled at the simpler ones at the beginning - showing how different AI models may reason in different ways.

This experiment highlights an important reality: despite their linguistic and computational power, AI models may not yet rival humans in abstract, non-verbal reasoning - especially under time pressure.

Can you beat AI on a test of pure logic? Take the same official Culture-Fair IQ Test and find out.

Find out more about the experiment at https://worldwide-iq-test.com/articles/chatgpt-iq.