I’ve noticed that the “Stop” button in ChatGPT no longer works as it used to. Previously, I could stop a response and immediately refine my message. However, now:
Clicking “Stop” always results in an error, forcing me to regenerate before I can edit or clarify my query.
Editing my message while ChatGPT is responding disables the “Send” button, preventing real-time adjustments.
This change significantly disrupts my workflow, as I can no longer quickly iterate on my queries. The previous behavior allowed for faster, more dynamic interactions, and its removal has made using ChatGPT much less efficient.
I strongly request that OpenAI restore the ability to stop a response without triggering an error, allowing users to refine and resend messages immediately.
Thank you for your time and for considering this request!
I have noticed the same thing. I agree about restoring that ability.
As it is the interactions between chats within a project no longer being available is hard for me but then when I realize that I need to add the entire context from previous in the chat or in another chat within the project the error messages and often even having to reload the chat definitely disrupts work flow and time.
I also hope the OpenAI Team restores the raw function of the Stop Button – stop generating the answer and let users input again.
Now, even if users make typos or find that the answer being generated is not what they want, they have to wait quite some time for it to finish. It’s inefficient.
I don’t know what you are seeing, but I believe this belongs under bugs.
Does yours:
Hit stop to cancel his message (important to reduce wasted tokens once you have your answer)
Then gives an error, offers to regenerate,
Option A) Hit regenerate, continuing to waste those tokens. Unable to cancel message.
Option B) Refresh page, page is now bugged, cannot send messages, stop button permanently displayed, not able to interact with it, chat destroyed? Come back later? Restart browser/PC, clear cache maybe? I’ll see if there’s a solution but to me it clearly appears to be demonstrating unintended behavior.
Believe it or not, it’s actually a good thing that it’s erroring out… I think… If it’s doing what I think it’s doing, it’s not actually registering the response as a used response when you hit stop. Which means you still get that response in your limit. If you have say just a standard account I’m not entirely sure though. Wish one of the devs would respond to that
You know you could just add something at the end of your question instead of In the next prompt saying can you make it longer at the end of your initial response you add something like
exact, couldn’t describe it better,
I like ‘editing’ or improving my prompts on the fly by stopping, since this update (?) I’m wasting a huge amount of time, and isn’t ChatGPT wasting an enormous amount of resources that every answer needs to be fully generated while no one is really waiting for it or wanting it ?
Why the h$ll was this ever reconsidered as better ?
I agree, this is so bad for my workflow, hope the devs will restore it. Now when I mess up my prompt I have to wait for the Wrong answer to appear. Or I have to abort, scroll to my last prompt and edit this.
I’ve found for me if I just re-fresh the browser page after I stop the run it shows the prompt input again, so the only diff in work from is a page refresh after stop, still stupid annoying but better than waiting for something I don’t want.
This is massively annoying. When I can tell that ChatGPT has misinterpreted my prompt, I don’t want to wait for it to finish before I can correct it. FIX THIS!
It’s getting to the point that I’m using other LLMs. The Stop functionality is what I loved about OpenAI’s ChatGPT. Until they fix this my workflows is brutal using ChatGPT.
The same all my devices. Seems this is the global problem. This is a super annoying, I’m using ChatGpt on my work full-time and impossible stopping is really spend a lot of time. Please fix it ASAP. Thanks you! Best regards
