It’s not even just slow. Sometimes when you’re watching a message scroll, it just crashes. The entire message PLUS your own question is erased when you reload the chat.
Yeah, just cancelled the subscription. Tested out Grok yesterday on the free version. Holy moly, the difference in instant response, context window, and how up to date it is.
Im truly amazed. Swapping to that for a month to test, until GPT has recovered.
Cancelled subscription as well — just super slow now ….
Probably will be asking for more money for more speed soon.
Same. you start a project and end up deleting most of it so chat goes fast which is a pointless exercise. what a waste of time… fix it or do one
Me too! This has been too slow lately, and I need to clean up the chats, which is very inconvenient and still not very helpful.
Same issue. It’s becoming impossible to work with ChatGPT
its just impossible to use chatgpt now, also cancelled. Even gemini is better nowadays
I’m thinking on doing the same. It’s sooooooooooooooooooo slowwww using Chatgpt nowadays even with my plus account. And Gemini is mindblowing quick even in reasoning mode…! And as we don’t have any answer for the support of OpenAI about this issue, I think I will be moving from chatgpt plus to Gemini…
I’m a Plus user and my analysis has been stuck on “paused” for 5 minutes. Honestly, after reading the comments here, I don’t feel like upgrading to Pro. I’d rather put my money into Gemini, even if some GPT features are better. 
These days I have to wait up to two minutes for the page to load (and it frequently resets randomly while I’m typing) before I can even get my prompt entered. Another two minutes to process and respond, during which time the page often crashes and resets. Lately there’s no response at all and I have to reenter my prompt. I took the same task to Gemini on the free tier and was able to complete my entire project with no wait time and no significant downgrade in performance. I don’t know why I pay for this. Probably will cancel if these lag issues persist.
I’ve found that performance feels more reliable if I treat long work as a series of shorter, intentional sessions rather than one massive chat.
For example, I keep important context in a small set of local notes/files, paste or upload them at the start of a new chat, do focused work for a while, then start a fresh chat once things get long or sluggish. That way the model stays responsive and I don’t lose continuity, because the context lives outside the chat instead of inside it.
It’s not a fix for backend issues, but it’s helped me avoid drift and slowdown when I’m working on something time-sensitive.
One small practical thing that’s helped me when it feels “stuck”: if a response just isn’t starting, I’ll close the tab, open a new one, and navigate back to ChatGPT. A lot of the time that alone gets things moving again.
It’s not a real fix, but it’s saved me a few times when things felt unresponsive.
I’m also a Plus user and have been experiencing similar. I find that the app is much faster at generating the responses than the website especially on longer chats. Really wish that OpenAI would put more emphasis on improving this.
I actually made a free little Chrome & Firefox extension to fix this in long chats, by rendering only the messages that are currently on screen - and rendering older / newer messages as you scroll up. It runs 100% locally, no tracking/data collection etc.
Its not perfect, but it helps a LOT making longer chats fast again, without losing any context.
you can try it by searching for “Speed Booster for ChatGPT“ in the Chrome / Firefox web store.
If you try it, let me know if it helps!