I don't know if it's related to the rerouting problems that have been reported in the last few days, as I've found [this thread from a month ago](https://www.reddit.com/r/ChatGPT/comments/1n478ta/deep_research_turned_to_useless_trash/) which seems to describe my issue. I'm a Plus user as well. One of my favorite use cases for Deep Research was to write longer-form fanfiction. As in, I would write a prompt that goes like "Write a 25000-word story set in so-and-so universe, where such-and-such happens, it should have an emotional tone, include this character...", and so on, and it would be able to write such a story in one shot, reflecting on itself for half an hour to figure out the details. It worked very well by my standards. Not perfect for sure, the writing did have minor issues, but overall it did produce coherent and entertaining stories with good prose, following my instructions, meaning I could generate this sort of on-demand fanfiction catered to my tastes, and I was very satisfied with it. I tried this again today after not doing it for a while, and immediately noticed it didn't work as it used to. In fact, the results were insultingly bad. In one instance, it tried to "write" a story by making a Python script that would combine a very limited bunch of sentence and name templates together, repeated over and over until it reaches my requested word count. I noticed that, as the month-old thread I linked to says, my remaining uses of Deep Research have not decreased. I had 25 left before (as I said, didn't use it for a while), and after several attempts I still have 25 left, despite having clearly selected the Deep Research option. I also noticed that usually, when I gave a prompt to Deep Research, it would ask me for some clarifications before actually starting, and then have a progress bar until it was finished, but that wasn't the case anymore. And finally, I saw that instead of my "Deep Research" remaining uses changing, it's my "Agent Mode" remaining uses that were being used up, despite me not selecting it. I tested it, and sure enough, whenever I start a "Deep Research", it actually consumes an "Agent Mode" use. I should mention that this switch seems to occur even in normal use cases, as in actually asking it to research a particular topic. I tested that out, and while it does try to carry out the research I asked for (with more success than writing a story), it still used "Agent Mode" without decreasing the "Deep Research" uses. Obviously I'm mad / disappointed about this and the recent events. I would be very bummed about losing this use case. If you know of any other AI Deep Research-like tool that can one-shot such prompts with good results, do let me know, but I was under the impression that GPT was the only model where this would work to this extent instead of being limited to its original goal of researching topics.