hi, two times in a row now, Deep Research has delivered only the last ~30% of each report. Both times, it went off and worked for 20ish minutes and then displayed only the tail end of its findings. Both times I then asked it to display the full report, and then to create a text file with the full report, and got this message: “Sections 1–5 were generated but not shown and are, regrettably, not currently retrievable from this interface.” I also created a new conversation for my 2nd attempt and still got the exact same result. One quirk: the first attempt produced a normal activity log from 32 sources as if there was a full report, but the second attempt produced no activity log and “0 sources” even though it still generated the last ~30% of a new report.
I’ve reported this to Support who didn’t offer anything that worked or any insights into the cause. And—and this part is really irking me—Support says there’s no mechanism to credit me back two Deep Research requests for these failed/corrupted attempts.
I don’t want to keep wasting my monthly allotment of DR requests without making any adjustments or trying something new, but I have no clue what to change. Really super grateful for any help / ideas / suggestions, thank you.
I have the same issue (chatgpt plus user).
Lost all my attempts and now I have to wait until end of month.
Please fix this bugs and offer the output as a well formatted pdf file.
Same problem here. I signed up for the expensive subscription specifically so I could get access to Deep Research. I understand that it’s a new application, but there should at least be some way of reimbursing folks for incomplete outputs. Can’t wait until the bugs are ironed out.
no matter what I try the results keep getting truncated. usually cutting off nearly the entire research response text, often only showing the citations at the end. I’ve tried everything.
Tried specifying a more reasonable response length, ie 2500 words limit.
Tried not specifying any word count at all
Tried every time including instructions to NEVER TRUNCATE the response.
Nothing worked.
I had a very promising initial 1 or 2 deep research responses, which got me very excited about the potential of it. Then, it never really worked for me again, so I had to revert back to plain old ChatGPT without the deep research enabled to “finish” what I was working on, although I really ended up just quitting on my project entirely at this point.
Same thing just happened to me. Seems like it randomly cuts off and only shows the end. Happened only with my last 2 requests I did today though. never had problems like 10-14 days ago, when I last used it. Maybe they changed something recently…
Here, Deep Research is experiencing breaks in the text, which are being hidden by references throughout the text; there are sections where sentences are incomplete, with huge gaps in the text.
