Am I crazy or has ChatGPT’s “thinking” time doubled this year while the answers stayed the same?

A ChatGPT Plus subscriber has observed a significant increase in the model's "thinking" time in 2025, noting that what previously took 3-5 seconds now often extends to 10-15+ seconds. Despite this longer processing duration, the user feels the answer quality has not proportionally improved, sometimes even seeming the same or slightly worse. The user questions whether this is due to heavier models, increased user numbers, or other factors.