Gemini conversation too long? Why it forgets at about 48 pages on the free plan, and the fix

How do you fix a Gemini conversation that is too long?

You move the work into a new chat with a handoff, because Gemini won't tell you a chat is full. On the free plan that happens after about 48 pages of text. There's one trap the other chat apps don't have. Gemini keeps answering past the limit, so the old chat may already have lost its first messages when you ask for the summary. The fix takes about three minutes:

1. Paste the handoff prompt below into the old chat and copy what Gemini writes.
2. Scroll to the top of the old chat. Check the handoff against your first few messages, and paste in any rule, name or number it missed.
3. Click New chat, paste the opener below with the handoff under it, and pick the same model you used before.
4. Read the three line reply before you ask for anything new.

```text

This chat is about to continue in a new conversation that cannot see this one. Write a handoff I can paste in as its first message. Use these headings and stay under 700 words:

1. Goal: what we are building or deciding, in two sentences.
2. Current state: the latest version of the work, quoted in full if it is under 300 words, otherwise its outline with every heading.
3. Decisions made: one line each, with the reason.
4. Rejected options: what we tried or ruled out and why, so it is not suggested again.
5. Rules: every instruction I gave on tone, format, length, names and numbers, including the ones from my first messages.
6. Facts: every number, name, date and file name we rely on, exactly as written.
7. Open questions, and the next step.

Add nothing we did not discuss. Mark anything you are unsure of with [check].

```

```text

Here is a handoff from an earlier chat that ran out of room. Read it, then reply with three lines only: the goal, the rules you will follow, and the next step. Wait for my go before doing anything else.

[paste the handoff here]

```

Step 2 is the one people skip. Gemini rereads the history on every turn, and Google says that once a chat passes the window its answers may not "take into account all the content provided". Your opening rules sit at the far end of that history. They're the first thing to go. So don't trust the handoff to carry them.

If only the last stretch of the chat went wrong, branch instead. On gemini.google.com, open the menu under an earlier response and choose to branch it into a new chat. Google's help page says deleting the branch doesn't delete the original, or the other way round. A branch keeps the history up to that point, though. It fixes a detour, not a full chat.

Does Gemini have a conversation length limit?

Yes. A Gemini conversation is limited by your plan's context window: 32K tokens without a plan, 128K on Google AI Plus and 1 million on AI Pro and AI Ultra, per Google's limits page read on 2 November 2026. Google calls the window "the reading capacity" and says a larger one "allows Gemini Apps to work with longer prompts and chats." Its upload help page adds the conversion. With 1 million tokens, Gemini "can understand up to 1,500 pages of text or 30,000 lines of code." We scaled that ratio down to the smaller plans. That's our math, not Google's.

Those pages include Gemini's replies, not just yours. Its answers usually run longer than your prompts, so a free chat fills up well before you've typed 48 pages yourself. Files count too. A 40 page PDF dropped into a free chat uses most of the window on its own, which is why Gemini's file upload limit matters even when the upload succeeds.

Deep Think is the odd row. Google's upload page says "If you use Deep Think, there is a 192 thousand token window", and its Deep Think page lists the feature for AI Ultra only. So turning it on in a 1 million token chat cuts the room by about 80%. And a long chat costs more than room. Google says Gemini has to "re-read the previous conversation history" on each turn, and it lists "the length of your chat" among the things its usage limits count. A long thread drains the meter faster, as Gemini usage limit reached explains.

What happens when a Gemini chat gets too long?

A Gemini chat that gets too long keeps going, and the oldest content drops out of what Gemini considers. There's no banner, no error and no warning. Google's limits page puts it in one line: "If you exceed the context window, this could lead to responses that don't take into account all the content provided". Its error codes page, read on 2 November 2026, lists more than a dozen codes and messages. None of them mentions length.

That makes Gemini the odd one out. We compared what each app's own documentation says happens at the edge of the window:

The ChatGPT wording is quoted on OpenAI's community forum, and ChatGPT conversation too long covers it. Claude's usage and length limits page says "When your conversation approaches the context window limit, Claude summarizes earlier messages", which Claude conversation too long walks through. Gemini does neither. So you have to spot the signs yourself:

• Gemini ignores a rule you set in the first few messages, like a word count or a name.
• It suggests an idea you already rejected.
• It contradicts a decision from earlier in the chat.
• Replies get slower, or pull in details that no longer apply.

The last one is the most common report. Android Police writer Anu Joy described it on 3 May 2026: "the responses slowed down, and Gemini started pulling in irrelevant details." When two of these show up, write the handoff. If a long chat throws "Something went wrong" while new chats work fine, that's a different problem, covered in Gemini something went wrong.

Why won't a long Gemini chat load or scroll back to the beginning?

A long Gemini chat often won't scroll to its first message because the web app loads older history in pieces, not because anything was deleted. In a Gemini Apps Community thread from 13 November 2025, a reply put it this way: "the conversation is too long, due to the way the interface loads older history. But your entire conversation is saved in your Google Account." That's a community answer, not a Google statement. It matches what people see, though.

Three routes get you back to the start:

1. Click inside the chat and press the Home key, or use Ctrl+F with a word from your first message. Both can force the older part to load.
2. Open myactivity.google.com/product/gemini. Google's activity help page says you can review your prompts there.
3. Export everything with Google Takeout. Click Deselect all, check My Activity, filter it to Gemini Apps only, and create the export. The Gemini box alone holds only your Gems, per Google's download help page. Google can email a link when it's ready.

Lag is the other symptom. In a February 2026 thread, a user reported that long chats auto scroll to the bottom on their own, and "It can take 1 to 2 minutes" before the screen settles. Pressing Esc or clicking into the text box is the workaround suggested there. Google hasn't published a fix as of November 2026.

Test it in ten seconds. Open a fresh chat in the same tab and send one short prompt. If it answers at normal speed, the long thread is the problem, not your connection. That's also your cue to write the handoff while the old chat still responds.

How do you keep long Gemini work from running out of room?

Start a new chat for each topic and keep your standing rules outside the chat. Google gives the first half of that advice itself: "Start a new chat when you move to a new topic." It says this also saves usage, because "Gemini won't have to re-read a long conversation history before creating a response."

For the rules, Google's personalization page says you can "set your own instructions" in Personal Intelligence, which lives in Gemini settings. Put tone, format and names there once. Then no handoff has to carry them.

Memory helps less than it sounds. Gemini can draw on past chats, per Google's memory help page, if you're 18 or over, signed in with a personal account, and have Keep Activity on. It doesn't work in Gems or Live chats, and Google warns it "might not always get it right." So a new chat may remember that you like short answers. Only a handoff reliably carries a draft.

Put long reference material in Gemini Notebook, not in the chat. Its help center says "The current limit is 500,000 words per source or up to 200MB for local uploads." That beats pasting a long document into a free chat, which holds about 48 pages at most.

Upgrading moves the wall a long way. AI Pro's 1 million tokens is about 31 times the free window. Room isn't recall, though. The 2023 Lost in the Middle study found accuracy "significantly degrades when models must access relevant information in the middle of long contexts", and what is a context window explains why. Even a 1M chat benefits from a handoff now and then.

In Gemini CLI you can steer the summary. Run /compress, which Google's command reference says will "Replace the entire chat context with a summary." The configuration docs set model.compressionThreshold to 0.5 by default, so compression starts at half the window. Use /chat save before a big change if you want a copy to go back to.

Whizi, which publishes this page, won't give a long thread more room. Every model on Whizi gets 40,000 input tokens per turn, and older history is trimmed silently, the same quiet behavior Gemini has. For one huge conversation, Gemini on AI Pro is the bigger window.