In the three years since the public launch of ChatGPT, we have witnessed a cognitive leap that would normally take a biological species millions of years. We have moved from AI that could “mimic” a graduate to AI that systematically outperforms PhD-level experts in complex reasoning.
Tracking the IQ Jump (2022–2025)
The progression from GPT-3.5 to the “Reasoning Models” of late 2025 (such as the o1 and o3 series) reveals a startling acceleration. Using benchmarks like the Wechsler Adult Intelligence Scale (WAIS), we can track this ascent:



