keen bisonReported by @chrome perch
#plus version, chatgpt 4o can't read imported stuff and generate images, o1 preview don't thinking
spare sleet
same
proud plume
same
timid badger
same
mystic iris
sam
clear lance
same 😦
wet holly
Same
vital geyser
Almost all the latest bug reports are about that same issue.
real edge
same
lapis field
same
mystic iris
same
little relic
same
patent gust
same
lone sonnet
openAI should consider removing dalle 3 and gpt-4o from free users, as their usage makes the service more expensive. they could provide gpt-4 and dalle 3 free access to only one Google account per user. If a free user logs out and signs into another google account, they should not be able to bypass the limits. thats openAI should implement a robust verification process for new users when registering or logging in. yhis will help prevent misuse of free offers and ensure fairness, protecting artists and other stakeholders from negative impacts.
thin bison
same, it's been a week 😦
tulip talon
The same here, usability is degrading significantly, it's terrible, it's not displaying images, it's not handling long texts, it's not analyzing anything, the intelligence has dropped too much, and performance is getting worse and worse, honestly, it's becoming a terrible service.
lime idol
samee
urban hearth
same
peak wind
same
wooden moat
same
urban chasm
same
wooden moat
the problem is simply that is using just the model that was on october 2023 so every new function are not working....
unique sparrow
Same. It would be really helpful if Chat GPT gave meaningful messages when hitting limits, throttling, or is otherwise broken. Just having it fall back to acting like it's completely unaware it can generate images, or suddenly responding as if the entire conversation was about a completely different topic... are pretty bad user experiences. It's MORE infuriating when the LLM acts like it can't do something than it would be if it simply said "oh, hey, servers are at capacity and I can't do that right now" or "huh, I didn't get a response from DALL-E, there might be a problem" or whatever. But don't gaslight me and act like you "don't have the capacity to generate images, but should you find a tool that can do that, here's a description of the image you described."
wheat light
same
errant grotto
mellow plaza
Same
rough cobalt
same
wheat light
What can we do te be listened? Tbh i am angry with this situation
lime idol
When openai solve the issue? Our membership is going. Issue has happening from 27th november and nobody contact or do something.
tired turret
same error here
rain crest
Hello,
I would like to report that the issue I was experiencing with ChatGPT (inability to read imported files and generate images, as described in my previous report) was resolved for me during the night of December 3 to December 4, 2024.
I should mention that I am a user in France, which might suggest that the issue was fixed progressively or based on regions.
Additionally, I don’t know if this is related, but my subscription was renewed and automatically charged on December 3, 2024, just before the issue was resolved. This might be worth looking into.
That said, I must point out that the resolution of this issue took quite a long time, and no official announcement was made to inform users of the fix. Moreover, no refund or compensation measures were offered to me throughout the duration of this error.
Thank you nonetheless for your work.
Best regards,
Lucas
wooden moat
The support team said to me that they fixed the problem but i still have it.…
lime idol
issue continue for me
blissful rose
issue continue
chrome perch
I understood what the problem is, right now, if I use a vpn in the USA, the chat cannot generate images and analyze them, and o1 does not think, but with a vpn in France, everything works for me so far. In some regions, it will most likely work, and in some it will not
Previously, changing the region did not change anything for me, but now everything works for me in some regions
fierce lodge
issue still continue
forest tiger
issue continue
vital geyser
Just now, I've done a test asking ChatGPT to create an image, something that was not working before. The image shows that in one response it gave me the image/logo i requested (was not doing that before) and the 2nd response is with the bug (not creating images).
So, they are, most probably, trying to fix it through feedback/choice or the 2nd is just the o1. Either way, now is working for some reason after 3-4 days with that bug.
chrome perch
Just now, I've done a test asking ChatGPT to create an image, something that was...
It seems like 2nd choice is o1, because it's thinking
zealous jackal
hello guys
chatgpt at the moment is unable to
browse online, generate images, and so on
on the BROWSER VERSION.
Is it like this for anybody else????????
please let me know that Im not the only one this is happening to
mighty minnow
its been a week man
faint scarab
wooden moat
Yeah a bit embarassing
chrome perch
Hello everyone, today I checked the work of chatgpt again, and this time it surprised me, not only did it bring back the old problem in all regions, but also began to imitate the "thinking" of the o1 version. I tell you, 4o, as before, does not read imported files and does not generate images, but o1 began to think and show it in the chat. But how did I know it wasn't o1? It's very simple, I logged into the chatgpt mobile app, which has always worked properly for me, and decided to check o1 there. And it turned out that this original o1 has learned to read user settings, as opposed to browser version, and starts from them at the deepest level. Conditionally, if his personality is registered in these settings, which he must adhere to, he will adhere to it so much that he will not even be able to say why, he will not be able to give out the text of these settings, which an ordinary 4o perceives as the very first message and passes its generation through the prism of these settings. Instead, when trying to find out from o1 why he behaves this way, in order to get information from him about these settings, that he took his identity from there, he breaks down and refuses to continue the dialogue, or blocks the message with a red warning. Chatgpt has been getting weirder lately
timid badger
the support team answered me in the same way.
this is written by a bot and it doesn't work(
I have some of the functionality back, but it works unstable during the day and may disappear
mighty minnow
the support team answered me in the same way.
this is written by a bot and it do...
So what should we do now? 😦
Does that mean no one can help us?
timid badger
So what should we do now? 😦
Does that mean no one can help us?
I found on the open ai website that they launched an update on November 22, and it will last 2 weeks, I hope everything will work soon
lime idol
Hi there, Thank you for providing a detailed report of the issues you're experiencing with the OpenAI web interface and the ChatGPT Plus functionalities. It seems like there are a few separate concerns here, so let's address them one by one: Authentication Token Issue: It's good to hear that half of the issue has been resolved and that the o1-preview model is working intermittently, and sometimes the o1 model is accessible. However, the persistent problem with the authentication token affecting access to certain models and functionalities is concerning. This issue might be related to a temporary service disruption or a specific bug in the web application's authentication handling. Our development team continuously works on improving the platform's stability and performance, and your detailed report is invaluable for identifying and resolving such issues. o1-mini Selector and Input Length Error: Regarding the o1-mini model selector showing the model as available but not processing inputs due to an "input too long" error, despite having worked with longer inputs before, this could be related to a temporary glitch or an update in the model's handling of input lengths.
We recommend trying shorter inputs as a temporary workaround while we investigate this issue further. Model Access and Drawing Capability: The intermittent access to models and the ability to use the drawing feature only after multiple attempts indicate there might be an underlying issue with model selection and feature access. This could be related to the initial authentication token issue or a separate concern affecting model functionality. Plus Account Upgrade Prompt: Seeing an "upgrade your plan" prompt on your Plus account is unusual, especially since your account is already on a Plus plan. This could be a UI glitch or an account status synchronization issue. For all the issues mentioned: Temporary Workarounds: You've already tried most of the common troubleshooting steps, such as using different browsers, the Windows app, and the mobile application. Since the mobile application is working fine, it suggests that the issue might be more related to the web interface or specific to the account's interaction with the web platform. Further Support: Since these issues are affecting your ability to use the service fully, especially as a Plus subscriber, we'll need to escalate this to our technical team for a closer look. Given the complexity and the potential impact on your user experience, resolving this is our priority. Could you please provide us with any additional details about the times and frequency of these issues? Specifically, any patterns you've noticed or specific error messages (besides the ones already mentioned) could be helpful. Rest assured, we are here to support you and work towards a resolution. We appreciate your patience and understanding as we work through these issues. Your detailed feedback is crucial for us to improve the platform, and we're committed to resolving your concerns as swiftly as possible. Best,
OpenAI Team
they return me this
mighty minnow
It is better to contribute in the OpenAI community forum here
https://community.openai.com/t/gpts-unable-to-read-and-process-pdfs-in-messages-anymore/653669
and upvote the post
atomic basin
Same
solar ravine
the same issue continues until now, no image recognition, not accepting any type of file input, no image generation and web search
hasty lagoon
Try disabling the memory, that fixed it for me.
timid badger
please note that when you click log out from all devices in the settings, this does not happen. this exits only on the device on which i clicked exit😆
thin bison
Try disabling the memory, that fixed it for me.
not working for me, but maybe it will work for other people
wooden moat
ITS NOW 2 weeks wake up people
eager timber
still no update?
crystal whale
Why is it that in the free version of GPT, when I upload an image to it, it can process it, but in the paid Plus version, when I upload an image to it, it can't process it?! The Plus version should have fewer restrictions, but it seems to be the opposite!
real edge
this is really too much, i'm paying for something i don't get
sick sierra
it is still not wroking for me is there any solution
thin bison
my case is fixed after i contacted them on https://help.openai.com/en/ by clicking the bottom right icon. start messaging them by clicking "chatgpt", "other", "other", "no", and "report a bug or errors". you can report this bug and it'll be resolved in a few days. hope this helps 🙂
mystic iris
same