Issue with Custom GPT Not Reading Uploaded Documents Properly

Description:
I am experiencing a critical issue with my custom GPT, which previously worked correctly but is now unable to retrieve information from uploaded documents.

Issue Details:

  • The GPT is no longer able to find specific terms and information that I know for certain exist in the documents.
  • Broader search terms, alternative phrasing, and manual verification have all been attempted without success.
  • The same documents, when uploaded and searched in a standard ChatGPT conversation (outside the custom GPT), return the expected results.
  • I even created a completely new custom GPT and re-uploaded the documents, but the problem persists.
  • The issue appears to have started recently, as the GPT was able to retrieve this information correctly last week.

Steps Taken to Troubleshoot:

  1. Verified that the documents were correctly uploaded.
  2. Attempted different search terms and manual review.
  3. Created a new custom GPT and re-uploaded the files.
  4. Compared results with a standard ChatGPT session (where searches work fine).

It seems like there is a problem with how custom GPTs process or index uploaded documents.

2 Likes

I am also having the same issue, the issue has started from 11-Mar-2025. In my GPT I have a large JSON file (3MB) and it struggles to get the data if the data is after certain number of lines (e.g. after 10k lines). It was working perfectly well prior to the 11-Mar-2025. I have raised a ticket with help desk.

3 Likes

Having the same issue, was working fine, now its stuck reading only the first page. Its driving me mad.

1 Like

Any update on this issue? I suddenly have the same issue. Everything was working properly until yesterday. I am based in Japan

Might be the same issue that I have; The AI model used is 4-turbo instead of 4o.

Try this prompt:
What is your ai model? Can you check without checking internet?

same thing happen but i use vanilla gpt it keep happening

Having the same issue. Seems my Custom GPT’s only read the first few pages in my knowledge files. It is causing mayor issues as they read the Table of content of my files but not the actual content. This leads to a lot of hallucinations as the GPT knows that it should be able to answer but does not referene the actual content and therefore makes a guess..

1 Like

So, I have managed to salvage somethings from my project. Basically, once your files goes above 12000 words in projects, chatgpt will hallucinate about that file, they will only read about 2000 words, then fill in the rest with hallucination. Other files in the same projects are able to be read properly(unless the files are supposed to be integrated together)

If you keep your files under 12000 words, you might be able to do work.

Anyway, I don’t know the exact numbers or the exact rules, what worked for me might not work for you. Anyway, you should file a support ticket

Also, this is most likely caused by the lates update on 3/27. The length of context is reduced. They might fix it, they might not.

1 Like