Competition in the global artificial intelligence (AI) industry is heating up as reasoning capabilities rapidly advance across models from tech giants like OpenAI, Google, and xAI. Latest data from TrackingAI, based on the Mensa Norway IQ Test, reveals a significant leap in the intelligence of these AI models in 2026.
Test results place models like Grok-4.20 Expert Mode and GPT-5.4 Pro at the top, achieving IQ scores of 145. This figure significantly exceeds the average human IQ of around 100, placing both models within the intellectual standard of 'gifted' or genius in abstract reasoning tests.
The assessment involved 35 questions focusing on visual pattern recognition, symbols, and relational logic. Interestingly, models equipped with vision capabilities tended to outperform pure text-based models due to their ability to process visual patterns directly, much like human sight.
This increase in scores occurred rapidly, considering that in 2025 the highest AI score was still at 135. However, experts warn that an IQ score is merely one instrument for measuring logical ability. This parameter does not cover other crucial aspects such as creativity, ethics, programming proficiency, or contextual accuracy in natural language processing.
Therefore, the title of the smartest chatbot remains subjective for users. The best choice of chatbot should ideally be tailored to specific needs—ranging from work productivity and complex data analysis to creative content development—as each model possesses distinct characteristic advantages beyond mere IQ numbers.
