This worked for me, following moribust and AC_developers steps.
detail": "File stream access denied
I had this problem for a long time and I managed to fix it within Google Chrome, by going to the three dots at the top right → Settings, and at the very bottom where it says “About Chrome.” Mine was outdated and it updated automatically there. I restarted Chrome and now I can download files without any issues.
Code interpreter session expired
I asked ChatGPT to put files I wanted in a zip folder & it solved problem.
I got around this error by prompting ChatGPT: “recreate the file in a way that generates a proper external download link”, then clicking the link downloaded my csv file.
02.12.2026 issue. Failed to get upload status for /mnt/data/filename
Hey guys,
Not sure if this exactly helps, but I pciked up a pretty consistent pattern when it comes to download files made in the chat that get the Failed to get upload status for /mnt/data/*filename* error.
So often when I got this error, I tried to hard refresh the page and it seems to show that there was never actually a response, more often that not showing “Stopped thinking…” despite me never trying to stop it and then sometimes just saying there is an error trying to get the message (Forgive me I can’t remember exactly).
After I refreshed to see this and re-tried, it seemed that the next pass was as good to go and instantly downloaded the files.
Again I cannot confidently confirm that this is a real and actual solution as it doesn’t always work, but it certainly helped save my sanity a bit.
Please let me know if anyone finds the true reason behind this and how to fix it!
EDIT: I just had another issue where I stayed on the chat, it showed me the full pass I asked for with the link to download the file etc, I refreshed the page and asked it to try again and it returned this message.
FYI
After the last post is an AI button that will create a Topic Summary.
I apologize if someone is looking for a more technical answer, or something involving coding that works outside of GPT, but I was getting the same error within CGPT and I simply asked what happened and all it had to do was generate a new link for me. Easy peasy.
`
I’m running into what looks like the same underlying issue.
Older generated attachments from the same chat still download correctly through /estuary/content:
https://chatgpt.com/backend-api/estuary/content?id=<file_id>&fn=<stored_filename>.zip&cd=attachment&ts=<timestamp_or_token>&p=fs&cid=<content_id>&sig=<signature>&v=<version>
But the latest generated attachment was only available through /interpreter/download:
https://chatgpt.com/backend-api/conversation/<conversation_id>/interpreter/download?message_id=<message_id>&sandbox_path=%2Fmnt%2Fdata%2F<generated_filename>.zip
After the chat reached the maximum length limit, the interpreter-backed link stopped working (probably deleted), while the older /estuary/content links still worked.
It looks like the final generated file was never promoted to durable file storage, so it was permanently lost when the interpreter session expired or was deleted.
This is quite annoying bug, can someone on the OpenAI team look into this?