Content windows and Other issues

To the OpenAI Product and Development Teams,

I’d like to raise a concern regarding some recent changes to ChatGPT, particularly the increasingly restrictive conversation limits, memory limitations, and the overall experience for users on the Free plan.

First, I’d like to say that I genuinely appreciate what OpenAI has created. ChatGPT has become an incredibly useful tool for millions of people, myself included. Its ability to understand different subjects, maintain conversations, assist with creative projects, and provide useful information is something I haven’t been able to find elsewhere to the same degree.

However, recent changes have made the experience increasingly frustrating.

1. Conversation and Context Window Restrictions

One of my biggest concerns is the new “Context Window Full” issue.

Conversations that previously could continue for extended periods are now being abruptly stopped once their context limit is reached.

This is particularly problematic for people who use ChatGPT for:

- Long-running roleplays and creative storytelling.

- Writing projects and world-building.

- Business planning and management.

- Organizing finances, projects, and important information.

- Ongoing personal conversations and emotional support.

I understand that conversations have technical limitations and that running AI models requires substantial computing resources.

However, completely stopping a conversation rather than allowing it to continue through summarization or another form of context management significantly reduces the usefulness of the platform.

2. The Previous Memory-Full Problem

Approximately a month ago, users experienced another frustrating issue involving ChatGPT’s memory limitations.

Free users were reporting that conversations could become blocked after only a handful of messages because of memory-related restrictions.

Now, we’re experiencing another situation where conversations are being cut off because of context limitations.

Having two separate problems interfere with basic conversation continuity within such a short period is concerning.

It also raises questions about whether these changes are being adequately tested before affecting the user experience.

3. Accessibility for Free Users

I understand that OpenAI offers paid subscriptions for access to more advanced models, larger usage allowances, and additional capabilities.

I have absolutely no issue with that business model.

I don’t expect unlimited access to the most advanced AI models without paying.

However, I believe there should be a distinction between limiting access to premium capabilities and limiting basic conversational functionality.

For example, someone starting a small business might rely on ChatGPT to help organize their business plan, track expenses, develop marketing strategies, and manage projects.

They might not have the financial resources to afford a subscription yet. That’s potentially one of the reasons they’re using the free service.

Similarly, someone who uses ChatGPT for creative writing or long-running roleplays shouldn’t have to constantly abandon conversations and reconstruct their stories because the context window has filled up.

People also use ChatGPT for ongoing personal conversations and emotional support. Consistency and continuity can matter considerably in those situations.

I believe Free users should still have access to a reasonably continuous and dependable conversational experience, even if they’re using a basic model.

4. My Suggestions for Improvement

I’d like to propose a few potential improvements:

- Allow conversations to continue when their context windows become full, using automatic summarization or older-message compression.

- Provide more reliable conversation continuity for Free users.

- Reconsider hard conversation cutoffs that prevent users from continuing their existing conversations.

- Maintain reasonable usage restrictions without making basic conversational functionality feel like a premium-only feature.

- Improve testing and communication before implementing changes that significantly affect existing users.

Paid subscriptions could still offer larger context windows, higher usage limits, advanced models, and additional features.

There is no reason that improving the basic experience must eliminate the value of paid subscriptions.

5. Final Thoughts

I want to emphasize that this feedback isn’t coming from someone who dislikes ChatGPT or wants everything available for free.

Quite the opposite.

I enjoy using ChatGPT. I appreciate its capabilities, its knowledge, and its ability to participate in creative projects and meaningful conversations.

I’ve tried other AI platforms, but I haven’t found another that offers the same combination of knowledge, accuracy, conversational ability, and creative flexibility that I enjoy here.

That’s precisely why these recent changes are disappointing.

I don’t want to constantly switch platforms, restart conversations, or spend more time managing the AI than actually using it.

I would genuinely appreciate it if this feedback could be forwarded to the appropriate product development team for consideration.

I also hope OpenAI will consider the broader impact these restrictions have on its users, particularly those who rely on the platform for creative projects, business assistance, and ongoing personal conversations.

A basic AI model should still provide a dependable basic conversational experience. Users shouldn’t have to purchase a subscription simply to avoid having their conversations abruptly terminated.

Thank you for taking the time to read my feedback.

I hope this reaches someone who can genuinely consider these concerns.

Buying a subscription doesn’t even guarantee the issue will resolve. Paid users have reported experiencing this as well.

Not surprised. The team really needs to step up their shit. They are going to lose subscribers if this persists. And since that’s their main source of income, they will lose overall money to support the AI.

If they do say this is a intended feature, they are going to lose a lot of reputation and drive away users from potentially doing subscriptions.

Not a good move. I get it. They need money, but this is definitely not a good way in attracting new subscriptions. Not to mention last months incident was also doing the same thing cutting off conversations after just a few inputs.

So they got two options, Fix it asap or admit it’s a new feature and everyone’s going to blow up at them and lose everyone who is subscribed and push away potential subscriptions.

Idiotic team sometimes. Just hand the whole AI’s to Elon Musk at this point, he will buy it and y’all can get ur money and he can fix the ai and improve it ten folds without any issues and balanced availability.

Okay so well I’ve been facing the same “context window full” issue and since I use it for creative work it really downgraded the whole experience because I’ve had to create like 7-8 different chats for one chapter.

Anyways, I contacted the support and the responses, if summed up pretty much say that this is “an expected behaviour”

I will attach the three responses I got for reference below -

Response 1 :

"ChatGPT Free recently changed how it handles conversations that reach their context limit. Previously, a chat could keep going while older content was left out of what ChatGPT used to answer. The newer behavior can stop further messages and display “Context window full.” This can explain a sudden interruption even when your usage hasn’t changed.

The current behavior is described here: https://help.openai.com/en/articles/20001639-why-does-chatgpt-say-context-window-full-or-chat-memory-full

If this doesn’t help with your issue, please respond to this message.

Best,

OpenAI AI Support Agent


Response 2 -

Thank you for following up with OpenAI Support.

We understand that your concern is not simply that conversations have a context limit, but that you noticed a significant change in how quickly your conversations reach that limit compared with your previous experience.

As outlined in our Help Center article, Why does ChatGPT say “Context window full” or “Chat memory full”?, “Context window full” means the current conversation has reached the amount of information ChatGPT can actively process within that chat.

We also understand that you have already tested multiple new conversations, including conversations without automatic summaries and on the web, and that you continue to encounter the message after approximately 12–20 messages. We do not need you to repeat those same tests.

Regarding your questions, the current behavior can prevent additional messages once the conversation reaches its available context limit. We’re unable to manually restore the previous behavior where a conversation could continue indefinitely while older content dropped out of active context, or guarantee a specific number of messages before the limit is reached, since context usage depends on the amount of information contained in the conversation rather than message count alone.

For now, the supported approach when a conversation reaches its context limit is to continue in a new conversation and carry forward the information you need, such as through a summary. We recognize that this does not fully preserve the natural continuity of a long-running conversation, especially for creative writing and ongoing discussions.

We appreciate you clearly explaining how this change has affected the way you use ChatGPT.

Best,

Christian

OpenAI Support


Response 3-

Thank you for reaching out to OpenAI Support.

We appreciate you following up and understand why maintaining continuity in longer conversations is important to you, particularly for creative writing and ongoing discussions.

We’ll take your comments about improving conversational continuity and the experience when a context limit is reached as valuable product feedback.

Regarding whether improvements are currently being considered or whether this will remain the long-term experience, we’re unable to confirm or provide a timeline for future product changes that have not been announced. For now, the current documented behavior remains as described in our Help Center article, Why does ChatGPT say “Context window full” or “Chat memory full”?.

When a conversation reaches its available context limit, the supported approach is to continue in a new conversation and carry forward the information you need, such as through a summary. We recognize that this may not fully preserve the continuity and smaller details of a long-running conversation.

Thank you for taking the time to explain how this experience affects your use of ChatGPT. Your feedback has been documented.

Best,

Christian

OpenAI Support


So yea in short, this behaviour is intended and … We don’t know if they are gonna fix it or keep it this way

We’ll share your suggestion to summarize or compress older context automatically so Free users can keep the same conversation going. The point about testing and communicating disruptive changes is clear, too. No timeline to share for those changes.