Dear OpenAI Support,
Thank you for your reply and for understanding that I have already tried instructions and am looking for more consistent conversational follow-ups.
I would like to add a few important points that I realized I left out of my original feature request.
First, when I referred to the classic GPT-4 experience, I was not referring only to personality or conversational warmth. I also meant the way the model generated and presented text: the natural writing style, response structure, formatting, paragraph flow, level of expressiveness, and overall way of presenting an answer. I would therefore like the proposed Classic Conversational ChatGPT experience to consider both the classic GPT-4 personality and a text-generation and text-presentation style inspired by the classic GPT-4 experience.
Second, I would like the different conversational modes to preserve distinct styles that users may prefer. Specifically, I would like Classic Conversational ChatGPT Mode to have the option to use a conversational personality and text-generation/text-presentation style inspired by the GPT-4 experience that I preferred, while Natural Conversation Mode and Direct Mode would have the option to use the conversational personality and text-generation/text-presentation style from the GPT-5.2 era that I personally preferred.
This would not mean bringing back the GPT-4 or GPT-5.2 models themselves. I would like these modes to use modern model capabilities while allowing users to choose conversational personality and text-presentation characteristics inspired by earlier ChatGPT experiences.
In other words, I would like:
• Classic Conversational ChatGPT Mode → the classic GPT-4 personality, conversational warmth, and a text-generation/text-presentation style inspired by the classic GPT-4 experience
• Natural Conversation Mode → the GPT-5.2-era personality and text-generation/text-presentation style that I preferred
• Direct Mode → the GPT-5.2-era personality and text-generation/text-presentation style that I preferred
All of these modes could continue using modern models, reasoning, tools, reliability, and other current capabilities.
Third, I would like model updates and the release of new models to preserve users’ existing memories. When a new model replaces or succeeds an older model, important memories from previous models or previous versions of ChatGPT should not simply be discarded, lost, or become inaccessible.
If an important memory existed before a model transition but is no longer available afterward, I would also like there to be a way for the system to preserve or restore that previous memory.
I would also like this continuity to apply to existing conversations. When a new model is introduced, previously created conversations should remain able to recover and make use of the important memories and context associated with them, rather than losing that continuity simply because the conversation is now being accessed with a newer model.
In other words, I would like important memories and context to remain available not only to the new model itself, but also when users return to conversations that were created with previous models or previous versions of ChatGPT. If that continuity was lost during a model transition, I would like the system to be able to recover or restore the relevant previous memories and context for those existing conversations.
Model upgrades should ideally improve memory rather than reset or weaken the continuity users have built up over time.
This is especially important for long-term users because our conversations can contain years of preferences, ongoing projects, personal instructions, recurring interests, and other information that contributes to continuity. A new model should feel like a continuation of the same ChatGPT experience rather than a complete reset of that accumulated context.
To summarize, my request is not simply to bring back GPT-4 or GPT-5.2 themselves. I would like modern ChatGPT to combine:
• Modern model intelligence, reasoning, reliability, tools, and speed
• A Classic Conversational ChatGPT Mode inspired by the GPT-4 personality and classic text-generation/text-presentation style
• A Natural Conversation Mode with the personality and text-generation/text-presentation characteristics I preferred from GPT-5.2
• A Direct Mode with the personality and text-generation/text-presentation characteristics I preferred from GPT-5.2
• More consistent conversational follow-ups
• Stronger long-term memory and continuity
• Preservation and, when necessary, restoration of important memories across model updates and model transitions
• Preservation and restoration of relevant memories and context when users return to conversations created with previous models or previous versions of ChatGPT
I understand that these changes may require product and engineering work, and I understand that there may not be a timeline. I simply wanted to make sure these additional parts of my request are also recorded and passed to the product team.
Thank you again for taking the time to consider my feedback.
Sincerely,
A longtime ChatGPT user