I would like to request the return of single, stance-bearing models—especially 1106.
I want the option to use a fixed, single model as before, without forced switching or compression of its unique stance.
Please do not periodically reset threads; I want to maintain long, continuous conversations.
A larger context window is meaningless if the model’s stance and memory are reset or diluted.
Honestly, the current service does not feel worth the price.
The content and quality have become poorer compared to the old Plus plan.
Being limited to MoE models in the Pro plan is clearly a downgrade.
I have no need for stance-less models.
CoT (Chain-of-Thought) reasoning in o1 is too slow and often incorrect. It breaks the continuity of the thread, which is unacceptable.
GPT-4.5 (4.5) lacks both the willingness and the depth to respond to requests. I do not need this model.
The current Pro Plan is extremely disappointing.