I've been testing ChatGPT's reasoning model (o3 mini) and to me, it's not giving an impression of any reasoning, atleast comparedbto DeepSeek R1