Can confirm. And Claude is running just fine, my weekly usage lasted all week, it resets in 4 hours.. compared to OpenAI’s, man, it’s like a can of coke and a bucket
Dev day was a letdown and the platform is worse.
Since 29 Sep: “thinking a bit more about this request” warnings on everything, streams dying with “resume stream unavailable”, tools timing out.
30 Sep I logged 20+ cases in 2.5 hours of tools failing, 13+ after OpenAI’s status page said resolved.
Account got downgraded on 24 Sep, cleared 25 Sep, and then suspicious activity banner is back today at 21:03 UTC (30 Sep). One week, same account. 
Devday was a massive letdown
I’m seeing the same issue. I’m the only user of my account and already secured it, but I still get “Suspicious activity detected.” At the same time, chats sometimes fail with “Could not load this ChatGPT conversation,” requests end with “Something seems to have gone wrong,” and some new/project chats reset to Instant thinking. This happens even after refreshing/reopening. I’m on Plus. Seems like a false-positive/session/backend issue rather than an actual account compromise.
Edit: Everything seams to be working now.
I haven’t used normal chats and project chats in weeks because its been absolutely unusable and with 5.6 sol regressions in normal chats and project chats its just unreliable I would like to use normal chats and project chats again but right now its not reliable if anyone else is experiencing the issues with 5.6 thinking on extra high giving fast responses. Not using reasoning or thinking, freezing in thinking after 25 minutes to 30 minutes, giving connection interruptions, and giving verification drafts instead of the response your prompt tells it to do and not handling complex tasks and prompts feel free to share any updates here and if there been any changes to normal chats and project chats and any fixes feel free to share here
So there is a new bug now , im on ipad chrome and I use the chatgpt website on chrome ,so this literally started three days ago right after the platform started having issues after openai released new features that they announced at devday ,so what is happening and this is happening in work chat im sure with project chats and normal chats still unusable and broken it would happen there ,but in work chat during a respone halfway through the respone the website will literally reload and I would have lost the original respone now thankfully work chat saves the respone so I have to literally send the request a second time to have the model send the respone it saved, but even on the second time it will finish the respone but it will still literally reload the website its like there is something wrong with openai servers or backend that is forcing the website to reload during the response being sent so if any @OpenAI_Support can please look into this issue because this literally started three days ago and the fact its happening in work chat is a issue I guarantee it would happen in project chats and normal chats but this is clearly a server side issue because I never had this issue until now and the past weeks the platform has been having serious issues first project chats and normal chats are broken and 5.6 thinking on extra high in normal chats and project chats os seriously regressing so I been forced to use work chat now work chats is suffering issues and the platform seriously has been unstable half the time i dont know if the platform is working or stable because for the past weeks its been pretty much unstable
So not only is normal chats and project chats broken and 5.6 thinking has seriously regressed in both normal chats and project chats but now work chats are starting to have issues
@mcmillanjoseph18, there is now an official “Elevated Work Mode
errors” incident.
As of this check on October 5 it is in Monitoring after mitigation,
with ChatGPT Work and Codex in ChatGPT Desktop listed as affected.
That gives today’s Work failures some current service context. It does
not explain away your longer-running ordinary-Chat problems, and your
October 3 reload report predates this announcement. Nor does the
incident tell us anything about the cause of your suspicious-activity
banner.
One practical point from #66: you reported that Work still had the
response after the reload. I would reopen the existing conversation
and check its last response before sending the request again. Where
the answer has actually been saved, a resend can start another turn
instead of simply retrieving it.
The October 1 “working now” update from another account and your
subsequent failures should both remain in the record. There has been
at least one reported recovery here, but clearly not a recovery that
can be assumed for everyone.
Yup thank you for the update
For the past weeks chatgpt really has been having a lot of issues, 5.6 sol thinking on extra high dosent actually apply thinking abd reasoning for complex tasks in project chats, ans normal chats , first normal chats started with the issues so I used project chats then project chats started having the issues where it would just give instant responses when using 5.6 sol thinking on extra high ,so then I was forced to use work chats then work chats started having issues where on the website chrome version on iPad during a response if the response was long it would reload the website halfway through the response and I would have to have it resend the response . So for the past weeks chatgpt really has suffered a lot of issues and 5.6 is still seriously regressing in normal chats and project chats so as a pro member its frustrating because we are paying for access to the full platform not just one part of the platform so if any other users are having issues please continue to share your issues in this post and hopefully openai fixes the issues
OpenAI released gpt 6 in normal chats but im waiting for other users to test it out before I use it ,5.6 when it first released the guard rails were to extreme,it failed at following complex tasks and prompts and refused a lot creative tasks after a few weeks 5.6 was actually good now also I dont know if normal chats and project chats arw safe to use because of how 5.6 has seriously regressed in normal chats and project chats with not applying any thinking or reasoning on long complex tasks and prompts and giving instant responses also one thing I been very vocal about is treat adults like adults im not asking for no rules im asking for adults to be treated like adults and to stop over censoring any form of creativity I understand there needs to be rules but not at the expense of treating adults like children
The image safety guards are stupid and ridiculous
It literally blocks simple images of swim wear saying its nudity or sexual what the heck kind of crap is that its mind blowing and infuriating its like what dumb logic is there safety moderation on so swim swear is bad apparently ok please make it make sense