community-help

[midnight mica]

I had the 'Voice and screen sharing' option earlier disabled under the speech section in settings in the browser, and it said 'Sharing your screen and video while in voice mode has been disabled by your workspace administrator.' After contacting you and asking about that, now the whole option is removed.

[charred parcel]

oh right, yes that's a known bug.

[midnight mica]

Ok then, could you please fix that for my account as I am a pro subscriber and other Plus accounts already have the feature?

[charred parcel]

I don't work for OpenAI, so I can't do anything. OpenAI employees have gold usernames.

[midnight mica]

Can you tag someone who can help, so they can check the issue and help me fix it?

[charred parcel]

the best way to solve the issue is to use https://help.openai.com and lodge a request there. Replies aren't instant and can take days to be replied to.

[midnight mica]

Yes, this is the issue. They are just ignoring me there, that's why I am trying to find someone who can help here

[charred parcel]

they won't be ignoring you, it just takes time for them to reply

there are 400 million weekly users of ChatGPT and only limited support

[midnight mica]

Yesterday, I was chatting with a guy from support. It was almost 1 minute between chats, then he just dumped the chat and didn't answer anymore

[charred parcel]

try lodging a new support request then

[midnight mica]

It has been more than a week in the same situation. I talk to them, then they stop answering me, then a new guy comes, gives a response and then disappears. The whole thing is useless

[sharp ocean]

search online for someone with the same problem

[charred parcel]

I wish there was more I could do to help but all I can suggest is to persist with support

[midnight mica]

I did that at first, but couldn't find anything. The issue is that I had the option but it was disabled, and after I talked to them, it got removed. It is something on their side.

[midnight mica]

Thank you, you did your best. This was my last try here to find someone who could help from their team

[golden socket]

hi. can i still use the Reason Feature if i hit the Free Limit on GPT-4o?

Can i use the Reason Feature on GPT-3.5?

[charred parcel]

no

also GPT-3.5 is gone

[golden socket]

oh i was wrong. i meant o3-mini. can i still use the Reason feature or is it only available to GPT-4o?

[charred parcel]

GPT-4o

o3-mini already has reasoning built in

are you sure you didn't mean GPT-4o mini?

[golden socket]

oh and i mean GPT-4o mini. sorry if i included that much

[charred parcel]

GPT-4o mini doesn't have reasoning

[golden socket]

What Model it will Use if the Free Limit on GPT-4o is Reached?

[charred parcel]

GPT-4o mini

[golden socket]

because i see o3-mini. can i still use that Model?

[charred parcel]

I suppose so, I haven't any idea on what models free users can use

[rugged wing]

i just paid for a subscription via the app store, while using that account i still have chatgpt free

how do i fix this\

[rugged wing]

my plus subscription is not appearing

[midnight mica]

OpenAI's entire service is broken. As a Pro subscriber paying $250 monthly, I've been requesting help for weeks without response. They've ignored me both on their website help desk and on Discord. I'm considering asking for a refund and switching to Claude or Gemini.

You might want to do the same if you value peace of mind. They don't care about their customers, all that matters to them is getting your money and then dumping you

[rugged wing]

@midnight mica i will, have issued a problem with apple.

sucks a lot tbh

[tawny tartan]

hi! anyone know how to get the age verification link? the page is expired and now if i open chatgpt it don't open the banner

[trail cairn]

https://help.openai.com/en/articles/9981739-what-happens-after-i-submit-my-id-for-age-verification#h_72b060772d

[tawny tartan]

ok thank you very much!!!

[empty stag]

why am i getting this

[tawny tartan]

also i have a problem! i had an account with my iphone that was logged with my apple id with the Plus subscription but i deleted it (i didn't know that is impossible to reactivate an account deleted) and now i did a new account with a new email with outlook. I payed a new Plus subscription with my iphone with the appstore but since this is logged with the apple id that is linked with my deleted account they covered me the money but i can't use the Plus and even if in my apple settings the subription results Active, in the app it results the free plan! i've contacted the open AI help and opened a coversation with an operator via email, but they said that i need to do another apple id or purchase the plus subscription via web... but i don't want to pay another subscription! Anyone had a similar problem or know how better manage this problem? they aren't very explicative :c

[trail cairn]

request a refund from openai or apple

[trail cairn]

it says connection error

[empty stag]

const fs = require('fs');
const { OpenAI } = require('openai');
const openai = new OpenAI({ apiKey: 'mykey' });

async function convert(file) {
  try {
    const translate = await openai.audio.transcriptions.create({
      file: fs.createReadStream(file),
      model: 'whisper-1'
    })
    console.log(translate)
  } catch (error) {
    console.error(error)
  }
}

convert('0310.mp3')```
thats the code do you know any solutions?

[trail cairn]

your ip is blocked or something on your computer is blocking it. does https://api.openai.com/ open from there

[empty stag]

yuh

[trail cairn]

can your node env fetch the same url

[weary tulip]

has anyone had any luck with chat creating a slideshow? It created the pitch deck (contents) fine but getting nothing when I ask about turning that into a slide show

[ancient herald]

I get “Unable to generate edits” error when trying to use Mac app direct edits code with VSC with o1 pro. Anyone else?
Other models work fine

[charred parcel]

is it only with o1 Pro that you have this issue?

[ancient herald]

Haven’t tried regular o1… but works fine with gpt 4.5

[charred parcel]

try with o1 and see what happens. I am curious to know if it's limited to o1 Pro

[ancient herald]

yep. o1 works. Pro doesn't

[charred parcel]

and you have been able to reproduce that behaviour a few times?

[ancient herald]

unfortunately, always.

[charred parcel]

ok I suggest to make a post in #1070006915414900886 because that shouldn't be happening

[ancient herald]

Done, thanks for the suggestion

[agile gale]

OK how do I solve this bug?

I had 11 prompts and now they're all gone

[charred parcel]

those screenshots don't provide any meaningful context, what is happening? what are you trying to convey?

[agile gale]

on the right, I had 11 beginning prompts in this convo but when I added to the first prompt in the conversation, I couldn't access the other 10 no matter what I did

So then I edited the first prompt but instead of it saying 12 out of 12, it went 2 out of 2

[charred parcel]

that indicator is just showing the replies after regeneration, not how many prompts you have left

[agile gale]

All the other ones were gone

But why delete them?! Is there a glitch?

[charred parcel]

I don't know, reach out to https://help.openai.com for more information

[copper willow]

does ChatGPT "sleeps" and "return as another guy" when not used for a long time? I use chatgpt and work at the same time. Minutes ago, it works good, it answers great. moments later, in the same chat, i ask it the question on the same topic and it's like "oh what do you mean about that? did you mean... (something out of my expectation)"

really needed answer 🙏

[lavish plume]

Does chatgpt currently have a limited usage for premium user?

[charred parcel]

Do you mean Plus users?

[rustic marlin]

looks weird, have you tried refreshing the page, for good measure? or even clearing cache, that might help

if it where actually deleted, you might indeed need to ask the support for help

[lavish plume]

Yes i mean plus users

[charred parcel]

yes Plus users have limits on the number of messages they can send to each model

[lavish plume]

Where can i find these limits, I didnt find it online

[charred parcel]

some limits are on this page https://openai.com/chatgpt/pricing/#free

the limits vary per model. For example GPT-4o is 80 messages per 3 hours and GPT-4.5 is 50 messages per week

[lavish plume]

Thank you! 50 messages per week is crazy.

[upbeat marsh]

Hello ! Do you know if there’s a way to extend a 5sec video created with Sora? I can’t find any « extend » button on the website 👀

[timber shadow]

Hey! Have you tried the "re-cut" feature? That lets you extend length to up to 20sec depending on quality, but since you can do this in chunks, you could in theory extend the end of one vid into a longer vid, then download both and load into basic video editing software, match up the matching frames, and extend indefinitely

[upbeat marsh]

@timber shadow do you have a tutorial or a video that explains how to do it? I don’t why but I can’t add a scene to create more that 5sec

[timber shadow]

In this video, after you click the "re-cut" button, you just need to change the length of the video from 5s to a longer length. That will extend your storyboard

The button between the quality (720p) and the # of variations (2v)

[upbeat marsh]

@timber shadow can’t clic on it 😅

[timber shadow]

If you change to 480p and/or lower the # of variations, can you select a longer length?

[kind rapids]

Discord Ops: I suggest adding a "GPT-4.5" tag to #1070006915414900886 and #1070006151938314300, as some reports might be specific to those models. I just had to tag a suggestion with "ChatGPT" but that's not the right tag. "GPT-4" might be suggested but to me that means the old/original 4.0. Thanks

[upbeat marsh]

@timber shadow Yes I can select 10sec if I switch to 480p…so that’s mean I can’t extend a 720p video 🥲
Thanks for your help !

[urban storm]

Hey all, a question, I have the teams subscription and I see the 4.5 AND the o3 mini deep research has limits per month? Its only 11 march and I allready reached the limit! Is there any way to get more?

[timber shadow]

Don't know if you already tried, but does it let you do 10s if you change 2v to 1v at 720p?

[charred parcel]

no, once you reach the limit you need to wait for it to reset

[urban storm]

That means I cannot continue to work for the rest of the month 😢

[charred parcel]

the only thing you could do is upgrade to Pro

[upbeat marsh]

@timber shadow no unfortunately the only way to extend my video is to switch to 480p, if I change from 2v to 1v I can’t clic the 10sec option 😶‍🌫️

[urban storm]

well the price of pro is a bit ridicilous

[charred parcel]

then wait for your limits to reset

[urban storm]

but honestly, a daily limit or a weekly limit makes more sense, a monthly limit,... 😮

[charred parcel]

GPT-4.5 is weekly

[kind rapids]

@urban storm you can use the API with any model. That's a pay-as-you-go option, pay-for-what-you-consume. That can carry you until the limit reset.
Now is a good time to recognize the value of switching models based on requirements. In a single session you can switch to a given model to make use of its specialty characteristics. This spreads the load across all limits. HTH

[urban storm]

oh, so I take the rest of the week off and then I can continue to work 🙂

[urban storm]

oh cool; thanks for the tip!

[kind rapids]

👍

[upbeat marsh]

So…I managed to extend the original video, how can I do now to extend more? (I don’t have the option to extend this 10sec video to a 20sec for example) 👀

[timber shadow]

Cool extension! This is where you'd probably need to start using some basic video editing software to keep going. Here's a basic outline of how that could work:

• Take your new 10s video and click "recut" on it
• Resize that 10s on the storyboard portion down to just the last frames (maybe 1-3 seconds? experiment with that) and move it all the way to the beginning of the storyboard
• This will make a new 10s video extending just from the end of that new video
• Now you'll have two videos: your orignal 10s video, and your new 10s video, and the END of the original video will be the same as the BEGINNING of the new video
• Since they have these frames in common, you can open both in some basic video editing software, line them up on the tracks where these frames are in common, and slice one or the other of the video clips, then you'll have a new, seamless longer video

[upbeat marsh]

Thanks a lot for your help! ❤️

[timber shadow]

You are welcome! #sora-discussions is a great additional resource - just shouting out that channel!

[upbeat marsh]

OMG this is amazing, I love the final result 🤩

[timber shadow]

Very nicely done! #sora-reels is a great channel to share your creations like this too!

[balmy cobalt]

Hi there. The CUA is available to all paid API users right?
because when I try to run the CUA sample app I get this error.

[rugged wing]

chatgpt plus subscription. I uploaded like up to 100 screenshots at once and it told me chatgpt 4o is limited and I need to upgrade to PRO SUBSCRIPTION. so is the pro the "TRUE" unlimited screenshot uploading?

[vivid mortar]

No, all models have context limits

[rugged wing]

What? So I can’t upload infinite screenshot messages even if I get PRO?

[vivid mortar]

Unfortunately Each model still has its limits, you would not be able to upload thousands of images as a single input into PRO, no.

Even pro has a context window/token limit, its just significantly more than non-pro plans

You can see some of the comparisons between plans on this page https://openai.com/chatgpt/pricing/

[rugged wing]

Ohh I see

[rugged wing]

How can I maximize the number of screenshot images I can send to ChatGPT then? Or is this uncontrollable since ChatGPT will take in what it needs to answer to my screenshot (for the context limits)

[vivid mortar]

Im curious what your use case is that requires so many screenshots within one chat.

If one conversation has become too large/long then I'd suggest just having it summarize the entire history of screenshots + conversation and move it into a new session with that summary

[rugged wing]

I don’t think my intended purpose is relevant to this conversation. I just get confused because sometimes I hit the limit quickly while other times I don’t. And apparently the limit is independent of device? I hit the limit on my phone preventing me to upload photos but I was able to upload screenshots on my iPad? Moreover they had separate reset times for the 3 hour limit

[empty stag]

Do i need to pay so i can use openai.audio.transcriptions.creste``model: 'whisper-1'?

@rustic marlin

[rustic marlin]

yes

there are no free models on the API

[daring granite]

I want to translate english to hebrew using gpt 4o model but it takes more than usual time to response, may be because of the large prompt which i am using but i have seen a website which uses this same model giving response in less time.

[mighty field]

Good afternoon!

Can someone help here with a bug? I am trying to use voice mode in the Chat GPT app on Windows 11, but this error appears.

The application has access to the microphone, in other programs the microphone works. I have not changed the settings of the app, uninstalled it and installed it again.

[ancient cove]

I've got a pretty infuriating issue — I'm being charged for a pro account with an account that I lost the email to. I forget the email, and after 2 massive threads with support, I'm still being charged. After days messaging in one thread, I was told they couldn't do anything, and I had to restart with a verified account. I don't see a way to do this: where is the log in area for openai supprt?

[charred parcel]

you can log in here https://auth.openai.com/log-in

you can access support at https://help.openai.com

[brave fractal]

Hi everyone, I'm thinking of finetuning GPT4o model to produce HQ news article following a natural article flow, vocab, and style of the author/news organization. Do you think It's possible with GPT4o? or i need to use opensource models like llama 3.3 70b?

If it's GPT4o, I'll be training it with 100M tokens, and for opensource model 1B tokens.
Dataset will be a prompt with style, site name, author name, facts, and expected output (that is the full article or paragraph)

[runic cove]

How to fix eror our system detect unusually activity from ur device🙏

[fresh flume]

I am struggling to get a custom GPT to follow any directions I give it. I'm not sure what to do differently, but it basically it just ignores all of its instructions and jumps into casual conversation and analysis when it's supposed to follow a very strict format and structure. As far as I can tell, I get roughly the same responses from a generic GPT 4o given the same data, meaning it's not using the custom instructions at all. Is this expected? Do I have to preface all of my instructions before I even engage the GPT? That's how it's appearing, which if so, this solution won't work. The goal here is to use ChatGPT Enterprise to assist in formatting technical escalations for my place of employment, which require a strict and rigid format.

[stable canopy]

Hi Everyone. I'm trying to roleplay with like immersive roleplay custom gpt... and after a while when the chat gets to long, its starts to lag and slow down, and i was thinking, that i could just copy the previous chat into docs and feed it the the gpt and it will know where we were in the prevous chat, but my chat is 270+ page long... the gpt can process that.... any idea how can i do this better?

[toxic falcon]

Hey, does anybody know why the API ai with the web search tool doesn't perform as well as chatgpt?

[toxic falcon]

you would basically need to summarize everything. keep out the small and umimportant details. give your document to chatgpt and ask it to fork and summarize everything in it

[stable canopy]

what is the max page chatgpt can handle? maybe i'll break down to multiple docs

[toxic falcon]

i sadly dont know. but 270+ pages for sure is way to long

[timber shadow]

Hello! One thing to keep in mind for this is the context limit of ChatGPT, which is measured in tokens. Once this context is filled up, older stuff is automatically forgotten (it leaves the context) to make room for newer stuff. Both your tokens and the response tokens count toward this limit.

Knowledge documents are not like normal context, either. The model accesses these by doing a fancy search, basically, where it takes your query and then tries to find info relevant to that query in the relevant doc(s), then return a relevant response based on what it finds (i.e., it does not contextualize the entire document[s] at once).

That said, you can definitely experiment with what you are describing and may likely even get desirable outputs, it just won't "know everything" all the time. But will well-crafted prompts, you can probably get the model to find relevant info in your docs, if you go that route!

[fluid kiln]

hi im new on dsc but i want ask about what hapend to o3-mini and 03-mini-high, for some time around 2-3 days i geting totaly ussles responses from chat, eaven in most basic question, comparing it to cloudie 3.5 (not to mention about 3.7) its like super down grade from open IA

is there some bug i dont know about

??

[stable canopy]

thank you tho ^^ chatgpt said 30-50 pages... so maybe if i break down like 25 pages... and i put it into the database... maybe just maybe it works 😄

[toxic falcon]

Your welcome 👍

[fresh flume]

And the custom GPT magically started working again, it's now using its instructions. 🤷‍♂️

[mossy gulch]

Anyone notice issues with pro right now all of a sudden it seems to not take any time to think?

[rustic marlin]

could you provide an example?

AFAIK, everything is fine

[past socket]

when are we going to see ChatGPT 4.5 in plus users & free?

[charred parcel]

Plus users already have GPT-4.5

[fluid kiln]

this is quick example i can provide right now:
main problem is resonating models, i cant exactly explain why but for some odd reason feels of chat with o3 and other resonating models is kinda like you talking to someone with memory of gold fish (with means chat after 3 or 4 messages dont remember what where before those messages)
in my experience somewhere around Feb it starts, where is main difference? i give you quick example from time where i was learning basic in dotnet maui with ChatGPT befor Feb and after .
before o3 was able to generate fully functional maui app with just main functionality description and rly small number of fixes (i was starting in those two attempts with same prompt)
but after some time i ask chat for the same thing (same model "o3-min-high")
and not only he start doing app in windows forms instead of maui but also used **xamarin.forms ** using for some odd reason i understand that maui and xamarin is similar but i tells chat that we are doing app in maui and dont eaven use word like xamarin in chat
at the end code he generated was not only un-usable but have mixed windows forms with xamarin and maui. ofc i do some fixes and re-coded xaml statments to be that of maui, but thats not the point, i feel like chatgpt was halicinating about instruction i give him or eaven ignoring some of them

i asked some people on this disscord and some other on reddit and they too feel like chatgpt resonating models become dumper for some reason

there is eaven a thread on your website https://community.openai.com/t/why-is-nobody-talking-about-how-bad-the-new-chatgpt-update-is/1108486

as for now im dont know if want to use openai solution for chat bots right now comparing to cloudie 3.7 chatgpt turns to be completely useless (ofc thats just how i personally feal about chatgpt right now)

im using 3 to 4 diffrent models but o3-mini and o3-mini-high been my persona favorite when it comes out, but now its un-usable

thats why im asking if smt happend to chat

[charred parcel]

the thread you posted is about 4o, the problems you are talking about are o3-mini

[fluid kiln]

i have same issue as those people i dont personally used 4o but the issue is the same

[charred parcel]

what do you want done about it?

[fluid kiln]

i dont know, im asking is this is normal that i feel chatgpt dont remember things i tell him 3-4 messages ago

[charred parcel]

no, that's not normal and has not been my experience either

[fluid kiln]

so what can i do about that? to return my chat to normal

[charred parcel]

how long is the chat you are having problems in?

[fluid kiln]

somewhere around middle of February

[charred parcel]

yeah, but is it a very long chat?

[fluid kiln]

here is also someone with same time and same issue

[fluid kiln]

nope mb 20-25 message

[charred parcel]

is the chat in English?

[fluid kiln]

yep

[charred parcel]

care to share a link to it so I can review it?

[fluid kiln]

this is chat after: https://chatgpt.com/share/67d2e921-aaa0-8003-aca9-0f16054e0ffb
and as for chat before Feb sorry but i dont have that one

[dry bone]

Hello, I've been having a new issue for over 2 days now generating canvas for the generated answer. It used to work perfectly until it just stopped. In chat it said the document is generated and ready to be reviewed but it is nowhere to be seen nor added to the list of existing ones. I waited a day to regenerate the answer but the problem persists. what am i doing wrong?

[charred parcel]

what is the problem in that chat? The model is doing everything you ask of it

[charred parcel]

which model are you using?

[dry bone]

GPT-4o

[charred parcel]

is the chat in English?

[dry bone]

yes

[charred parcel]

care to share the link so I can review it?

[dry bone]

hemm... i'd love to but not sure how

is part of a long thread

[charred parcel]

how long?

[dry bone]

fairly long ... don't have anything to comapre it but i'm using to do some data analysis so it's been quite extensive

[charred parcel]

when you said you had been waiting for a day for it to regenerate an answer, you know that ChatGPT won't spend a long time to generate an answer? If ChatGPT told you that then it was hallucinating and it might have been hallucinating because your chat is very long

[dry bone]

i waited thinking it could be an overload, out of tokens, glitch or something related to that. The answers are generated in chat (when told to), just don't generate the canva as requested even though it says it did

[charred parcel]

I would start a new chat

[dry bone]

damn ... i was afraid that was the solution (i tried in a different one and it worked)

[charred parcel]

ask ChatGPT to summarise the chat you are having problems with then paste that into a new chat and see if that helps

[dry bone]

will do

thanks!

[charred parcel]

no worries

[fluid kiln]

problem is that, those reply is not as good as they used to be in past, model besicly mix maui with xamarin

[charred parcel]

guide the model back to what it should be doing

[fluid kiln]

that is exacly the point in past when i was using o3 when it comes out i dont have issue with something like that also when model got stuck the "guide" to correct solution been easier than now, but mb its just my feeling and nothing more

[charred parcel]

do you currently have working code?

[fluid kiln]

yep but not that one from chat, i fix that my self

he only started code and i must personally fix issues

[charred parcel]

well I would take the working code and start a new chat with it

[fluid kiln]

i come to same conclusion but still have filling that chat responses are kinda dumper than it used to be

[tame bobcat]

When using deep research, does it matter which model I have selected? (4o, o3-mini, etc.)

[blazing ridge]

Hey everyone, does anyone know how to send PDF files containing text and images to the OpenAI API?

[wild saffron]

Hi everyone. I need to teach my mother basic math from middle and high school. My math's knowledge is pretty advanced from my university, thus it's everything obvious to me when I try to speak with her about math, but it isn't for her, of course.

Can GPT make a PDF structured file from where to start until where I need? In general, can chatgpt create PDF file with the content it gives?

[trail cairn]

it does not create pdf files

[hidden vigil]

Is there a way to remove an attachment from a chat so that I can change the model to 4o mini? Or to export the memory from that chat into a new one?

[keen maple]

anyone been able to install CUA in a docker image on a WSL2 subsystem?

[upper hare]

Has anyone else seen an increase in the past week with chatGPT (all models) being lazy or not following instructions. I have been using the paid tier for six months, it always has some issue but now I feel like it is becoming useless.

[restive mural]

Hi! I asked to ChatGPT to convert normal maps into displacement maps (different kinds of image, I am not sure that iis actually importnt). Atnyway, a the biginning, everythings worked correctly, I could convert and download the images without any issues. But since some days (I don't remember when exactly), ChatGPT dosn't give me the converted images anymore: he tells me he converted them and he will send them to me, but he doesn't, he ends-up his answer without giving to me the converted image... I told him several times that the image doesn't appear, he answered me that he will send it again but he doesn't... I sent him back the image that I want him to convert, but it didn't fix the issue... Could someone help me please?
Thank you in advance!

[vale prawn]

has anyone had an issue saying you've hit your billing limit even though there's plenty of credits in the account?

i even made a new key, i'm the sole user of the account

[restive mural]

I am not using the paid version and I am not sure axactly when it started, but yeah, since a while, I feel a drop in performance of ChatGPT, as well concerning its answers, its ability "keeping memory" about the conversation, etc... I have also more and more difficulties to upload an image and to make ChatGPT to do what I want with it...

[upper hare]

The biggest issue is when it get stuck it "wrong loops" where it will admit its mistake and then keep doing it. Even in you make ever more specfic instructions.

[unreal hemlock]

<:book_icon:1171408210398289941> Rule 8: Other AI discussions.

-# Discussions of non-OpenAI products/models should be posted and discussed only in the #ai-discussions channel.

[cunning river]

excuse me mr. Rule8, but that product supports OpenAI

[wintry nacelle]

I haven't been able to log into my account all day. Does anyone have an idea what's been going on?

[thorny crown]

do they not count prompt tokens in the api anymore? like when i prompt it? i dont even see it on the usage dashboard. the system prompt is at least 1k tokens but in 3 chats it just counts the users message not the system prompt

[restive mural]

Oh yes, it happened to me too... 😭

[potent stream]

I always fail using curl api responses, for chat completions it works. Curl api responses are successfully posted but the output structure of the openai documentation is not clear

[eternal seal]

Can the operator handle tasks? I told it to ask me something at a specific time and then complete Task X with the information I provide. It said it understood and set up the task, but I’m not receiving any notifications in the browser, app, or iPhone, as I would normally expect from tasks.

[solid harness]

I paid for the chatgpt pro account. I then created an API key to try and use it, but I get "Error: You exceeded your current quota..." how do I find out what my quota is and when it resets? I went to the API dashboard and it doesn't see any activity at all

[charred parcel]

ChatGPT Pro is for ChatGPT, the API is separate and billed separately

[solid harness]

Ok, so is there a way I can see my quota/usage/limits and when they will reset for ChatGPT Pro?

[charred parcel]

On Pro GPT-4.5, GPT-4o, o1, o3-mini, o3-mini-high are all unlimited, there is no quota

Deep Research is limited to 120 per month

o1 Pro mode is 100 per day

[solid harness]

Ok, maybe I was on o1 Pro or something, I'm not sure, it told me I was downgraded to a lower model because of use. So in order to use the API, I have to also pay? That is... kind of annoying

considering the app I am using to chat would use the API

[charred parcel]

the API is separate

the app you use for Pro is ChatGPT on the website or on a Desktop App

[solid harness]

I tried to install the ChatGPT android app, but it didn't work. So I tried another 3rd party version, and it needs an API key

[charred parcel]

the API is separate and don't trust apps where you enter your own API key

[solid harness]

alright, i got that the API is separate, but if OpenAI's app is broken, then I can't use it on my phone, except through the browser

[charred parcel]

use the browser then

[solid harness]

i will, it just is limiting, i can't use voice

[charred parcel]

get an Android phone that is modern enough to support the app then

[solid harness]

My android phone is not that old lol

[charred parcel]

if it's not old then it will support the app

[solid harness]

You really think that a Pixel 8 is too old?

[charred parcel]

Pixel 8 will support the app, is your OS up to date?

[solid harness]

it is up-to-date

[charred parcel]

what happens when you try to use the ChatGPT app on it?

[alpine terrace]

can we set which gpt model to use in the custom gpt?

[coral basalt]

Hello everyone,
I have a quick question, please

So, I have 5 large ZIP folders containing a ton of text and documents
I'd like to create a GPT, an AI, or something else, that would work solely and/or specifically on this database to meet my needs
I also have the premium subscription at 20 euros per month

Because currently, when I give it all the folders one by one, I feel like there's a limit, that it can't handle everything
I especially feel like it can't analyze the .zip files very well, or only partially, and I'd like it to exploit all the details found in the documents (there's only text, in .docx)

Plz, do you have any other solutions to suggest other than continuing to discuss as I'm doing right now?

Thanks everyone !

[torn lynx]

Hello everyone,
I went through the openai agents guardrails docs
As per this I can halt the execution when guardrail agent is not happy with my output..
What I was wondering is, instead of halting, my use case is around masking some sensitive data using guardrails from output.
is this something possible right now, if yes can you point me to some reference example or document.

Thanks.

[upbeat marsh]

Hello! Do you know if it’s okay to prompt in French with Sora? 🇫🇷🥖🥐🍷

[trail cairn]

plus users have 32k context limit to work with so what you are trying to do is impossible. it can only pick relevant snippets to the context

[long pawn]

Hey, I think I have a stupid mistake. Regardless of what I write, these are not stupid things, just things. Hey, what's up, how are you? This means to me what kind of policy violations someone else has.

[trail cairn]

you cant bypass model refusals, also ignoring moderation flags and continuing with unsafe data may risk your api access

[coral basalt]

Oh no, too bad.
If I send them separately via message, will that change anything? Will it all end up being accumulated, or is it dead?

[trail cairn]

the 32k limit is per message sent and it may include chat history, custom instructions, memories, etc so its even less than that and remaining will be snippets from the files but not full files. works kinda like this https://help.openai.com/en/articles/8868588-retrieval-augmented-generation-rag-and-semantic-search-for-gpts

[charred wadi]

Does OpenAI increase the message limits per some hours for free users?
I feel like I could message GPT 4o for way much more than a few days ago (usually it's around 10-15 messages, but now it's more than 30)

[charred parcel]

OpenAI don't explicitly state Free user limits but I suspect it is dynamic based on load

[fading solar]

Hi can I post the videos that I make with Sora on YouTube. Just to show people what I can do?

[charred parcel]

yes, you can

[fading solar]

Thanks so much!

[thorny crown]

That just makes u pay for the app to open ai’s expensive api prices. Then gives them ur api key to use too

[coral basalt]

Okay...
Too bad, thank you very much.

Is there a fairly easy-to-access solution you could suggest, or is there none, and too bad for me?
(Another site or another method, for example...)

Thank you very much.

[peak fable]

yo why is chatgpt writing responses extremely slowly rn?, the chat isnt big at all, nor are my requests too complex, i never had this problem before.
how can i fix it, or do i just wait?

[charred parcel]

which model?

[peak fable]

lowkey i dont see a modle name anywhere, just chatgpt

oh wait its 4o

[charred parcel]

ok 4o shouldn't be slow

[peak fable]

it was still in the middle of typing

yeah it wasnt slow before at all

the convo isnt even big enough

also its writing smaller responses that usual

[upbeat marsh]

🙃

[jolly kelp]

uhh is anyone finding the the edit message is gone again???

[trail cairn]

yep

[jolly kelp]

are you also getting something like this??

[gritty meadow]

Hello does anyone know what the length limit for chats is based on? Before getting the "You've reached the maximum length for this conversation" message. I'm struggling to find an answer online.

[trail cairn]

its counted in a way that you cant estimate it visually

[gritty meadow]

Can you elaborate?

[trail cairn]

it not only includes your prompts and replies but all hidden messages, regenerates, memories, tool messages etc

[gritty meadow]

Is this described somewhere online? So if I edit message number 56 out of 100 and "branch out" from there, the other 44 messages are counted towards the limit?

[trail cairn]

dont know. someone tried it with multiple tests + data export long time ago when chatgpt was less complex than it is today. for the same reason they probably dont show any counter as its too hard

[jolly kelp]

uhh is the chatgpt server down rn??

[charred parcel]

no, it's up

[gritty meadow]

Ugh I see. And excluding that, is there any info what it's generally based on, like number of messages/number of tokens?

[jolly kelp]

uhhh I'm getting this

[peak fable]

i just got that too

[trail cairn]

it was not message count based but text length, database record size sum or something

[gusty anchor]

did the website die or is it a local problem? even using vpn doesn't do anything

[peak fable]

gateway timeout

theres no way i lost what i was writing for half an hour right?

[gritty meadow]

Huh, I see, thank you so much. I thought it was just about the context size of the current "branch" of the conversation so the 128k tokens. But I'm really new to this, idk if I understand even the context correctly.

[gusty anchor]

im crying for you ;-;

[trail cairn]

for context window management it just truncates older messages and does some other stuff. the size limit might be some technical issue/limit

[gritty meadow]

Aha, so I can exceed the 128k tokens but still keep writing in the chat, it just won't really "remember" everything?

[trail cairn]

yes. also in chatgpt the limit is 8k for free, 32k for plus and 128k pro

[gritty meadow]

Thank you! "ChatGPT" refers to their website/app, right? as opposed to the API

[trail cairn]

yep

[gritty meadow]

Ok, thank you so much for taking the time to explain all those stuff to me

[jolly kelp]

uhh is anyone facing something like this for the regenerate button 0/3??

[obsidian oracle]

Hi, I have the Android app version 1.2025.070, and the Advanced Voice feature isn't working for me. The bot simply doesn't respond at all. However, after I exit the voice mode, I can see that the bot did listen to me and responded in text form in the chat. I've been experiencing this issue since the introduction of Advanced Voice. I have a Plus subscription, so I assume I should have access to this feature. Does anyone know what might be causing this?

[civic stirrup]

Hello folks 😁. I have a account problem. I log in into my iOS app or via OpenAI . But it doesn’t give me my plus account only the gratis account. My chat is visible.

[quiet berry]

anyone still having issue with edit button on everything outside of projects?

[waxen canopy]

Is anyone else having an issue with alternate responses being labeled 0/X, and being unable to switch between them? I seem to have the issue mostly with 1o

[jolly kelp]

uhh is the issue with the edit message button resolved??

[rustic marlin]

I've seen a few comments about that issue so far

I have a strong educated guess that OAI is probably working on fixing that

that sort of UI bug usually gets fixed pretty quickily

at some point you will open the page and will not even remember there was a problem =P

[waxen canopy]

I hope so; the 0/X is new, but I've also had problems with the navigation vanishing entirely for a few months now in certain chats, both the thumbs up/down, copy, read, regenerate, and switch buttons on the bottom and the edit button on the prompt

Another thing it does is jump from one response to another without input, which makes it difficult-to-impossible to find a specific section, as well as the desktop app randomly switching itself to 4o-mini--again without input--and blanking the display panel

[heavy moss]

Anyone help me to finetine a Gemini or OLMo model. Is distilation is better approach? I have data in pdf files. Thanks

[steady lantern]

Hey there, I'm setting up the Assistant API to work in my app using cURL. I just wanted to double-check I'm doing everything in the right order based off of the documentation: https://platform.openai.com/docs/api-reference/assistants

First, I set up the initial code such as the authorization, content type, API key, and whatnot; I'm pretty certain this is all working because I do get the initial response with all the id information, and the user input system was working with the regular version of ChatGPT before I decided to switch to Assistants API.
Next, I start by creating the thread and run: https://platform.openai.com/docs/api-reference/runs/createThreadAndRun
Then, I create message by extracting the thread and run id and putting them together into new url like here: https://platform.openai.com/docs/api-reference/messages/createMessage
Next, I retrieve the run again using the id strings to create the final url: https://platform.openai.com/docs/api-reference/runs/getRun

But, instead of getting the AI's response to the user's input, it gives me a bunch of errors as seen below. Am I doing this in the wrong order or am I missing some steps?

[real radish]

What happened to the past pinned chats with ChatGPT??

I used to be able to go back to groups of chats and continue the dialog but now I cannot
these saved conversations were critical to saving time and to our workflow when working on ongoing projects or maintaining context!!

Community Ask
is there any way to get back past discussions other than reconstructing it from scratch

---

when I asked ChatGPT it replied
It looks like OpenAI recently changed the way chat history is organized, and pinned chats may no longer be available in the same way as before. You might still be able to find past conversations by scrolling through your chat history, but the feature that grouped or pinned certain chats for easy access appears to be removed.

If there's anything specific you need to revisit, I can try to help reconstruct past discussions based on what I remember from our conversations. Let me know how I can assist!

[charred parcel]

were there ever pinned chats? I don't recall seeing that feature

[real radish]

yes you could create it by selecting the 3 dots to the right of the conversation in the list of conversations on left

made it easy to pick up on a conversation where you left off

it was actually called projects

huge time saver

[cerulean silo]

Hi, just today i received a message saying "Suspicious activity detected
It looks like someone else may be using your ChatGPT account. Please secure your account to regain access to all features." I tried all the step but it still showing it

[charred parcel]

contact support at https://help.openai.com

[primal cargo]

Hey guys, does anyone know how to cancel Team Suscription and change to Plus without losing your information and chats?

[charred parcel]

you only needed to ask once in one channel, I have answered you here #openai-chatter message

[tepid remnant]

I recently experienced suspicious activity on my ChatGPT plus account (auto-switching to GPT-4o Mini + odd messages). I’ve reset my password & enabled 2FA a Logged out all devices. Has anyone else encountered this? Any additional steps I should take?

[charred parcel]

report any concerns to OpenAI Support at https://help.openai.com

[burnt elk]

ayo people, for some reason i have gotten suspicious activity detected, secure your account notification and now i am blocked from most of the models. how to get past it?

[charred parcel]

contact OpenAI Support at https://help.openai.com/

[burnt elk]

yea... thats not how it works

[onyx kraken]

We're all community members here, the very few who work for OpenAI have gold names #server-staff . But that help.openai.com, the bot bottom right, that is the tech support method and how I'd seek help if I were affected!

[noble vault]

Does anyone know the current AVM limit for plus users?

[tender aspen]

I can't see the option to edit my existing prompt anymore in ChatGPT interface. I have to start new thread every time as a temporary solution. Did they really stopped it or is it only me?

[vivid mortar]

You may have to mouse over the message to see the edit option, but you should still see it here. I don't believe you can edit the first message sent within a thread, but you should be able to edit all the rest of them.

[iron rivet]

Around 1 hour a day for Voice, ~15 minutes of which can be used for Video...
The limits are a bit dynamic, so on lighter days it may get a bit over 1 hour (Like 1 hour and 15 minutes)

[supple steeple]

i had this problem in the app but it was resokved after a recent update

[junior marlin]

Hello lovely people 🙂 Where is a good place to ask a custom GPT question?

[timber shadow]

Hey! You can definitely ask your question here!

[junior marlin]

Thank you 🙂 I was wondering about attaching PDF's to knowledge in custom GPT's. Is it better to break up long PDF's? What are the limits?

[timber shadow]

The individual file limits are 512MB/2mil tokens, so if you have a file above either of those, you may need to split it up. But no matter how you split it up, the text is represented in embeddings, and the file(s) can have their contents searched via those embeddings. https://help.openai.com/en/articles/8843948-knowledge-in-gpts

[junior marlin]

Thank you that is very helpful. I see it says "The parser can struggle with multi-column PDFs, and won’t understand the nuance conveyed by the relative positions of text on a PowerPoint slide." To confirm does this mean if I say what is the last sentence in Chapter 23 it won't know the answer but it is still processing the information or is it specific to formatting?

[timber shadow]

To me that note sounds more like "it can process plain text as you'd expect, but if the spacing/physical layout of that text is important, that part won't be processed". I also don't know how it would do with a request like recreating a specific sentence, because accessing info via embeddings is all about searching for relevant info in a document based on a specific query, not "knowing" the entire document all at once.

[junior marlin]

Ok I will look into that further. Thank you so much for your help it's much appreciated 🙂

[vagrant mauve]

Hola, friends. Can someone point me towards the channel for feature requests? mine has to do with Canvas, if that helps.

[charred parcel]

#1070006151938314300

[vagrant mauve]

You're the best, thanks!

[tacit hare]

Hi

Someone have the same problem? detail
:
"Speech.create() got an unexpected keyword argument 'instructions'"
error
:
"Error generando audio". My code: params = {
"model": "gpt-4o-mini-tts",
"voice": line['voice'],
"input": line['text'],
"response_format": "mp3"
}

        # Añadir las instrucciones solo si el modelo es gpt-4o-mini-tts
        if line['instruction'] and params["model"] == "gpt-4o-mini-tts":
            params["instructions"] = line['instruction']

[mystic hill]

Hey, im currently making a project with gpt for fivem my fivem server. The peaple that are police are everytime vdm, always as a swat team. So everyone want a ai cop script. I have made that but it doesnt work.

The problem is they dont spawn.

[timber shadow]

@knotty bobcat Hey! If you don't see a login button at the help page, your login from ChatGPT probably carried over. When you start a new support chat, what options do you see?

[knotty bobcat]

I will just show you

[timber shadow]

It looks like something isn't loading right, in the middle Messages tab, there's an "ask a question" button that appears that lets me start a new chat.

[knotty bobcat]

continuing on

[timber shadow]

Ah ok since the login now button is appearing, it doesn't look like you're signed in currently. Try closing this page, then go to platform.openai.com and sign in there

Then return to help, you should be signed in

Also have you already done the easy steps: change your password + click the button to sign out all other sessions in ChatGPT?

[knotty bobcat]

now we wait

[knotty bobcat]

Yes I have

[timber shadow]

The chat history will be completely new once you sign in and return to help.openai.com, so it won't sign you in mid-chat I don't believe. That's why I recommended closing help.openai.com first, then sign in at platform.openai.com, then return to help. When the "login now" button isn't present at the start of a new chat, you'll know it worked (and you'll know that the support ticket is tied to your account)

[knotty bobcat]

I saw the typing indicator

]

popping up and off again

[junior marlin]

Hi there 🙂
I am setting up some academic research custom bots.
I was wondering if the code interpreter would help to be switched on with large attached text PDFs in a custom bot's "Knowledge"? Or does the code interpreter have no impact on the back end of the custom GPT bot? I just want to make sure I am setting up everything for optimal results.
Thank you for your time and patience 🙂

[rustic marlin]

cod eitnerpreter is just one of the tools the AI can call

[remote tulip]

What does Signed on this date of does not match pattern: ^(0[1-9]|[12][0-9]|3[01])-(0[1-9]|1[0,1,2])-(20)\d{2}$ mean? Removal request

[rustic marlin]

it does not directly have an influence over the behavior of the agent, but the provided context of the PDF may have a difference in how the AI acts depending on how the user interacts with it

for example, if you ask it "What is the average population of Foo country?".. it would probably either trigger a web search or hallucinate it

but lets say you have an table on the pdf with population over months...
then it might instead trigger code interpreter and calculate the value to answer the question

[rustic marlin]

where are you seeing that?

guessing by the regex pattern, I bet you are wrriting the date in dd-mm-yyyy, and the pattern should have been mm-dd-yyyy

[junior marlin]

Thank you so much that is super helpful. Have an awesome weekend 😊

[fading herald]

Hello, I have a question: Is it possible to change the login method? Currently, I log in using Google account authentication, but I would like to switch to using an email address and password instead. I do not want to cancel my current account and create a new one, so could you please tell me how to change the login method while keeping my current account?

[astral pulsar]

Unfortunately not@fading herald

[fading herald]

Thank you for your response. However, the Google account I use at work is different from the one I use at home, which makes it inconvenient. I hope this will be improved in the future.

[iron basin]

how do i use chatgbt screen sharing on my computer

[junior marlin]

Hi there! I have a question about something that I’ve been curious about for ages. Why does Chat GPT always spell Midjourney wrong? I know general predictive text also does it. So is it just a training issue? If yes do you think it will eventually be corrected? Thank you for taking the time to answer all these questions 🙂

[wide hatch]

any updates on this ?

[knotty bobcat]

It worked the next day, but then stopped working again at 6pm AEST time

which is right now

which is so annoying

[wide hatch]

going through the same situation here

[orchid pivot]

hey so i was just wondering why i can't access any other model even tho im on plus, i already do what the cs say, enable the multi factor authentication and log out.. , but it still give me the warning

i checked on the device that login into that account and i see no device other than mine

the only reason that it may or mostlikely raise the flag is because im using different vpn for remote work and college

but sadly even tho im using home network without even changing any ip or using vpn, it's still give me the warning

[knotty bobcat]

Welcome to the club

[manic basin]

Hey! I just asked 4o to add on to a piece of text, and it gave this incredible result! Is there A way I could get ChatGPT (or a different AI, like Clause 3.7 or Grok or Gemini or all) to work directly in Google Docs for free? (I don't mean to get access to an AI I'd normally have to pay for for free, but it'd be nice for an AI to see my footnotes & formatting) Maybe a browser extension?

[orchid pivot]

i usually just use canvas within the 4o and the 4.5

but if you'd like to try to integrate ur docs with gpt u could search for an extension on doc itself

dm me and i'll give u the link

[manic basin]

Alr! just did

[blissful ravine]

In Trae claude sonnet 3.7 keeps posting truncated modifications to files and getting stuck on the fact that they weren't finished before, any way to fix it?

[blissful ravine]

Unfortunately the official chatgpt extension sucks apparently

[amber juniper]

I had to change my phone number for work. The old one is still linked with my account. Is there really no way to change that?
I'm scared that it will be needed in some kind of situation at some point and i might lose access to my account. (the current setup i have and my projects is really important + its linked to my main email address)

[forest cipher]

Does anyone know how I can I use the chatgpt app on a degoogled phone? I have a Fairphone 4 with Murena OS.

[charred parcel]

you can't

[charred parcel]

for which model?

I use the macOS app and it's showing reasoning for me

[hybrid elk]

I am currently enrolled for 11 million free tokens for sharing prompts with openai, but i have just notices that my account balance has gone negative and api requests are being denied. the dashborad shows that i only used 2 million tokens today. has anyone expereience a similar problem?

[trail cairn]

used 2m on what model?

[hybrid elk]

4.5 preview

[trail cairn]

you only get 1m for that per day and rest is on your expense, the other 10m is for mini models

[solid hemlock]

With a plus subscription, can you be limited to 4o requests?

[charred parcel]

Plus gets 80 messages per 3 hours for GPT-4o

[charred parcel]

did that answer your question?

[solid hemlock]

Yes, thank you. But can the restriction be activated if I haven't reached that limit?

[charred parcel]

you should get 80 messages every 3 hours with Plus on GPT-4o, if you are getting less than something is going wrong

[shadow horizon]

Why do I keep getting this even when i use a new chat?

[rustic marlin]

try refreshing the page, it will not loose the history

it could jsut be that somehow that sesstion went into an invalid state and the browser and the server are not doing well with eachother, a refresh usually solve that problem

ctrl + F5, actually, not just a regular refresh

[shadow horizon]

I tried it and it didnt work :/

It just says now "error in message stream"

[rustic marlin]

are you able to start a new chat?

try completely closing chatgpt's tab, oppening a fresh one and navigating to this convo again

[shadow horizon]

now its just stuck like that

[unkempt cloak]

i confirm this is happening to me as well

[shadow horizon]

Guess we just gotta wait then

Tbf sometimes it works, just takes a lot longer than usual

[charred parcel]

keep an eye on https://status.openai.com

[shadow horizon]

Ohhh thanks :)

[green goblet]

Is anyone else getting constant network failures on the MacOS ChatGPT app? I get it as soon as I've made one request to o3-mini-high. I have to click "Retry" constantly and it's a big pain.

[charred parcel]

working fine here, I just prompted o3-mini-high and it responded to me with an answer

https://chatgpt.com/share/67e1891a-bfb8-800a-97bc-e348731d2af2

[green goblet]

It seems to go wrong on the second or third request.

[charred parcel]

i'll try a few more, stand by

working fine, I can't break it

[green goblet]

I wonder if it's my slightly crappy internet at times, and that the app can't handle it very well. But at the same time, it's mostly with the o3-mini-high model that it happens.

[charred parcel]

dunno, I can't fault it

[green goblet]

thanks for trying!

[rustic marlin]

just a while ago there was a issue on the status tacker, but it was solved

it is possible that it may take a while for everything to be deployed.. or the issue eprsists and we will probably see another incident on the status tracker soon

[green goblet]

I've had the problem on macOS for a number of weeks.

[rustic marlin]

in that case, it is for sure something entirely different

[green goblet]

Here's an example. Just happened.

[rustic marlin]

be sure to be on a stable internet conection

avoid proxyes and VPN

[charred parcel]

ok, I can help there, that happens in the macOS app when the IP connection is disrupted

[green goblet]

I don't have the best internet, but I regularly use zoom, google meet and other apps and none displays a problem like this. The app must be super sensitive. It also only happens right away. Once things are streaming, everything works fine.

[charred parcel]

you may be dropping packets, the macOS app needs a stable connection without packet loss

[green goblet]

Is it using UDP so it's unreliable transit? It's not like this is a realtime application of streaming?

[charred parcel]

the macOS app is sensitive to packet loss, if you drop packets it's going to have issues

[green goblet]

Do you know where I could log an issue report? Hopefully it can be made a bit more resiliant 🙂

[charred parcel]

sounds like a problem to raise with your ISP

[green goblet]

Does OpenAI have any issue trackers that you know of? (my ISP would not recognize this as a problem -- spotify, Netflix, etc all work fine).

[charred parcel]

OpenAI can't provide feedback about your Internet connection. You need to solve that or you will continue to have issues. To be clear, this sounds like your Internet is the problem, not the macOS app

[green goblet]

I understand your point. In my opinion this isn't likely a problem with my ISP and so would like to file a report with OpenAI despite that. I assume from what you are saying they don't provide any support even though I am paying. (I know you don't work there, so it's not on you and thanks very much for your help.)

[charred parcel]

OpenAI will provide you support, all you need to do is go to https://help.openai.com and file a support request

However right now there is nothing I can see that is playing up except your Internet connection.

[green goblet]

Fair.

[charred parcel]

Make sure your macOS App is on the latest version, make sure macOS in on the latest version, disable any VPN or Proxy software, try running a speed test on your connection to see if that shows any issues

[green goblet]

I'm trying a mobile network to see if that changes anything. VPN and proxy software is off other than Cloudflare Warp which should not be affecting this app. Hmm.

[void topaz]

We've had a nonprofit Enterprise workspace on ChatGPT for over a year, it got deactivated yesterday as the credit card didn't have balance, I want to pay the outstanding invoice today but for me as the workspace owner it's grayed out in the UI and no way to access the workspace to pay. Anybody has any ideas on how I can pay and reactivate our workspace? Anybody had issues like this?

[charred parcel]

you need to contact your OpenAI Enterprise representative

[void topaz]

I don't think I have one. I've been personally to both OpenAI DevDays but never spoken to anybody at OpenAI outside those events.

We pay 96 EUR per month 4 licences

[charred parcel]

is your account an Enterprise account?

[void topaz]

ok, maybe it might be another type of a Teams account

[charred parcel]

oh you have a Teams account?

[void topaz]

I have Pro myself also

personally

[charred parcel]

contact OpenAI support at https://help.openai.com

[void topaz]

I've searched through the articles there and found this

But this does not work, don't see any actual human contact info on help.openai.com

And just got an email yesterday from OpenAI with a link to an invoice to pay but it is now void

[charred parcel]

use the chat bubble in the bottom right hand corner and ask for support

[upper dock]

is there a way to download my sora videos in a batch (and get the same rez version - the chrome extentions Ive tried just seem to get the crunched version and v poor quality. Should be around 10-12 mb not 800kb)
Any ideas, thanks

[charred parcel]

ask in #sora-discussions

[frozen frigate]

might need a little help from an experienced user, is there anyone that has 5 minutes for me please?

i'm trying to make chat-gpt put together 3 different docx files into a single pdf file and enhance the information inside them also, it's for a medicine test at my college, and i need that material to study, but seems like he can't understand properly and now has limited my possibility to chat even tho i have bought premium.

[livid cape]

i am getting emails from OpenAI for stuff i reported.. but i never reported anything?

[north shadow]

what is the limit on 4.5 again? was it 25 responses a week or something?

[normal gorge]

Is there a way to make chatgpt render equations always use $ and $$ while also rendering correctly??? This is really annoying that it always defaults to outdated ( and [. I can use custom instruction to correct it but then it won't render properly in the web UI.

[orchid pivot]

@knotty bobcat any follow up on ur end?

[onyx kraken]

LaTeX and KaTeX formatting for the web environment, windows, is something I've played a lot with.

It sounds like you may be thinking you correct it, but the path you picked ?maybe? isn't the one that the web environment for ChatGPT actually needs?

This is what I'm aware works, and of course this is just basics, but with this the model generally gets to fancier stuff (colors, grids, various other cool stuffs)

$$
E = mc^2
$$

% Inline LaTeX within a line:
\(E = mc^2\)```

https://chatgpt.com/share/67e238a7-e6c4-8011-b18e-f7914e4925dd

The second screenshot's just a paste of one I showed like 6 months ago from some other discussion about colored LaTeX and grids or something.  There's a lot the model can do with this.

Does this help?  If I didn't understand, please clarify.  This is a topic I love to explore and share.

[onyx kraken]

Hi!

You kinda have a lot of questions. I am not sure what is most important to you, I will offer advice about the idea of 'taking 3 PDF, enhance the info, and making 1 docx you can download'.

ChatGPT is like a hammer. It is best if you tell it 1 thing to do at a time - or a few things in a clear, normal order, familiar stuff.

Yes, we can do fancy and unique stuff with it too, but I recommend get good with the basics, like driving a car - you don't start off your first drive trying to go 1000s of miles in bad conditions and all kinds of roads all at once. You learn how in small ways.

It sounds like trying to do the stuff in small ways, one step at a time, may help you.

You don't say how big the PDFs are. Really big ones can be hard for the model.

But the pathway I recommend, it have it see each PDF 1 at a time. Have it take from each PDF what you find useful (all of it, some, whatever - the model can only hold so much - it's like a fridge, not like the entire Stanford University Library - you try to put a LOT of stuff in it and it's gonna just hum like a confused fridge until you really know how to guide that)

Then show it the next PDF.

Then talk with it, get all the information 'set up right' so you are happy with it.

Then have the model make you the docx document and the download for it.

Hopefully then, success. Discuss if you have questions, hope I or someone can help!

Aha. I missed it's 3 different docx - into 1 PDF - just as good/even easier. But same steps

[thin osprey]

Looking at the Structured Output documents, i notice that the API is like: client.beta.chat.completions(). Is the "Beta" suggesting that it is not recommended for production usage?

[sonic escarp]

@rustic marlin

I have logged all out and changed my password but this has been for a while and wont let me get back to making or building it kept erroring

[rustic marlin]

from my experience with the API so far, the API have been changing quite a lot. OAI usually tries to keep backwards compatibility, but thye only do so with an announced date of the official deprecation of any given feature which may be something like.. 6 months.. 1 year...

[rustic marlin]

the error message is very clear. it has nothing to do with the content you are generating, it is just warning you about a possible issue with your account. follow the suggested steps to secure your account

[sonic escarp]

I did already 2 days ago

changed my password last night. still getting the errors

it keeps sending me codes to sign in and logged out of everything yet it doesnt let me do anything

[rustic marlin]

it might be a good idea to inquiry the support about that then. https://help.openai.com

[sonic escarp]

I cannot even get this to be right to show my PROOF OF WORK and that is so sad https://canary.discord.com/channels/974519864045756446/1353977119222530059

[sonic escarp]

I have

[rustic marlin]

by ay chance, do you see any convos in your history that you don't think you created?

[sonic escarp]

no responces other than what I have already done no aid

[sonic escarp]

no

all my own stuff it got mad at me for trying to code and learn a new skill

saying its not what I usually do. what I'm trying to get a new job it doesnt want me to learn and grow?

[rustic marlin]

the warning ahs absolutely nothing to do with the fact yo uasked it to generate code

[charred parcel]

OpenAI really needs to improve the UX with these security messages

[sonic escarp]

It is infuriating I just want to grow and learn

[rustic marlin]

it is because the system detected unusual activity on your account, such as it being used fro man IP address that comes from soemwhere like a different country...
this is very common in cases where your account credentials have been phished, changing the password is good, but that not always solve the problem if you have some sort of malware on your machine or a shady browser extension siphoning your session tokens

[rustic marlin]

agreed

[sonic escarp]

"Inconsistent Usage Patterns: Sudden spikes in activity or settings changes." is under the https://help.openai.com/en/articles/10471992-why-am-i-receiving-a-suspicious-activity-alert

yes I have been coding or trying to and it keeps breaking so activity spikes

[rustic marlin]

for now, the only thing you can do is to wait for the response from the support

[sonic escarp]

"Multiple Concurrent Sessions: More simultaneous logins than usual." YES BECUASE it keeps cutting off my ability to see things and having issues KYC

[rustic marlin]

there aren't many more troubleshooting steps that could be done on your side since it is something on the server side

[charred parcel]

what you need to concentrate on now is getting support to help

from what I have seen of this issue with multiple users, OpenAI are taking a "shoot first, ask questions later" approach to security and when users are securing their accounts making them wait days or a week for a response from support.

OpenAI need to do better and their current approach is unacceptable

[sonic escarp]

it trully is esp for even the minimum pay tier it is like we are fully lower class citizens

all their staff has is stock answers as well

[rugged wing]

hey, im experiencing constant crashes since the last update. i think it is the auto correct feature? as i may be writing and suddenly the app vanishes. does anyone know if this is a known issue?

[humble raft]

soooo i have a chat that has frozen and has data i need to recover for a project i am working on. i have refreshed the page. closed the browser and restarted. tried another browser... nothing seems to work the chat is just frozen and i can not send any inputs.

[trail cairn]

if it has reached its max limit its over. you can try to check the browser console for errors or adding a message from some other app to see if it gives that error

[humble raft]

other chats work fine

[trail cairn]

yep but single threads have a length limit

[humble raft]

oh this just popped up in it

Suspicious activity detected
It looks like someone else may be using your ChatGPT account. Please secure your account to regain access to all features. Learn more.

is this because i tred accdessing it from another browser?>

[trail cairn]

probably

[humble raft]

still frozen however

[gritty meadow]

I noticed that in one of my chats one of my older messages has disappeared (but chatgpt's responses before and after are still there). I wonder if anyone knows what could've caused this and how to avoid it in the future?

I noticed in the api response that this particular message has:

{
   "status":"in_progress",
   "metadata":{
      "finish_details":{
         "type":"interrupted"
      }
   }
}

[fresh flume]

Okay, what changed? I have had to re-provide the context 3 separate times in this conversation. That's a first since 4o.

[fresh flume]

Okay, perhaps it's related to the error rate? This is happening frequently enough I feel like I'm going to have to start checking the OpenAI status page before I even attempt to work with it.

Also, is that exclusively API? Because I'm using the app.

[hot brook]

Hello. I’m still confused about DeepResearch somewhat. I was able to give it a zip file with a compressed GitHub repo for analysis, and it said it could analyse it, but I just realised now that I can’t give it zip files if the normal model is set to o1 pro, for example, but it accepts the file if I set it to 4o. I thought DeepResearch was its own model and the model selection menu didn’t affect it? Or is it just ignoring the zip file? I just don’t want to feel like a fool thinking it’s using the files.

[rugged wing]

hmm

[vernal canopy]

Does it feel dumber like they rolled back past updates that made it know you better and give better first responses and better responses in general? Does it feel dumber?

[blissful sail]

is there any way to resize the brush tool in sora image generation?

[fresh flume]

I'm not sure. Response times are way faster than it was 2 weeks ago, but it also seems more error-prone. The responses are punchier.

I'm thinking something with context management has changed

[west silo]

Hello,

I'm facing a big issue with billing & invoicing.
OpenAI is charging me while i've deleted my credit card data from my account and canceled the subscription 3 month ago now.

OpenAI is literally restricting access to billing & past invoices to all users who are not subscribing to "Plus" which is not allowed in Europe especially in some countries (i'm from France)

The support is not helping at all.

[grave tundra]

Srsly, why does ChatGPT on o1 makes so much mistakes?!??!

I asked for assembly coding to avoid using t7 registers, so a specific command, and yet it still corrects my work while using t7 registers
And it does so many mistakes while asking simple prompts, asking to correct something to which I obviously don't have any answer on internet or anywhere, and yet it blocks and gives me something that just doesn't work.

Isn't o1 supposed to be really performant??

[normal gorge]

The tools that I use (Typora and 2 others) don't support "(" and "[". I can force chatgpt to generate latex formula with only "$" and "$$" but then it doesn't render correctly and just shows the raw formula.

[loud prism]

I bought chatgpt plus on my iphone but when log on pc it said i use free version does anyone experienced this?

[onyx kraken]

Oh! I never heard of this stuff 🙂

But I'm a community member like nearly all of us are, thanks for explaining! The very few of us who work for OpenAI have gold names.

I wonder if they're aware of the issue. We have a #1070006915414900886 and a #1070006151938314300 option, and I'm not sure where your issue would fit. If you're willing, maybe pick one and help make sure OpenAI knows about the issue and the need for a workaround for people like yourself?

[onyx kraken]

Potentially a cookies issue, maybe?

Myself, I might restart my comp and see if that makes everything better.

If not, tech support through https://help.openai.com/en/articles/6614161-how-can-i-contact-support log in if needed and click through the choices until you can report a bug

[loud prism]

Weird that my ipad and my phone can use 4o version but my pc can't thx i'll check it out

[west tree]

I am using Open AI, O3 model and getting streaming reponse but Token output is None, I want to fetch input, output tokens, How to resolve this? Please tell me if anyone know the solution?

[timber shadow]

@cosmic forge Hey! Do you have another device you can test on, or another browser you never use to test in?

I think the best test might be something like a web browser on a smartphone

[cosmic forge]

thanks for the help its not that big of a issue you know

[timber shadow]

Ah gotcha 😁 sounds good! Feel free to respond if you want to look into it again

[cosmic forge]

yes will do

[marble bridge]

hi. does anyone know if o1 pro, gpt 4.5, and the new imagegen are unlimited with the pro subscription?

[charred parcel]

o1 Pro is 100 messages per day, GPT-4.5 is unlimited, image generation is unlimited all on the Pro subscription

[viral trellis]

can anyone tell me why i am getting water mark in video generation?

i have chat gpt plus

can anyone tell me

[charred parcel]

you need Pro to remove the watermark

[viral trellis]

please

ooohhhhh

sad😂

i thought in plus i get watermark free😂

[charred parcel]

nope

[north nexus]

Guys someone said that GPT4o is free to use for all users but there's limit right?

[fresh sedge]

Any suggestions,i couldnot create videos in sora

Being a plus member

[upper briar]

does anybody know how to do this.
i'm currently on a free plan based in the uk

[charred parcel]

GPT-4o images is rolling out to Free users but isn't completed yet, you should be able to produce 3 images per day on the Free plan once it's rolled out.

[upper briar]

ahh. because on poe when i try to do it, it comes out like this

[charred parcel]

you can only generate GPT-4o images on ChatGPT or Sora.com

[upper briar]

for me, on sora it keeps saying "plan recquired"

[charred parcel]

yes you need to use chatgpt.com and have an account, you can use sora.com on a GPT Plus or Pro account

[upper briar]

doesn't show this new "4o" feature

[charred parcel]

as I said, it's rolling out, the rollout is not complete for Free users, you will have to wait. If you don't want to wait then sign up to a Plus subscription.

[upper briar]

okay

can i use paypal

[charred parcel]

no credit card only

you can sign up on iOS the ChatGPT App or on Android through Google Play Store

[shy sigil]

I've been playing around with the new image gen stuff and it's pretty fun. But can someone show me how to prompt properly for the best possible results like some of the posts on reddit?

[misty ember]

Sora keeps saying Server at capacity. Since last night

[rustic marlin]

it isn't really extracting anyhting, the input is embedded and it is just generating a completely new image that looks like the input, if you try to overlay both you will see how it doesn't really match

well.. for must purposes, it is probably good enough

[granite sigil]

if i buy pro will this fix it

or is it at capacity at capacity

i have plus but my generations are stuck in sora for about 10 minutes, is that normal? seemed much faster yesterday. how reliable is the speed if i upgrade to pro?

[charred parcel]

I have Pro and I just started a new generation without issues

720p 20 seconds

[woven stag]

how many hours left until office hours ?

[rustic marlin]

some people are reporting errors on image generation

we might see something posted on the status page soon..

[woven stag]

hey, new to this server
is there any info about the office hours event ?

[rustic marlin]

or.. it will just get back to work with no post about it too...

[rustic marlin]

https://discord.gg/openai?event=1354557954786988042

[brisk tulip]

i am trying to access sora.com and i am getting this message: Your session expired, please go back and login again. When i go back and login, it takes me to the same page.

[charred parcel]

clear all your cookies and your cache, restart your browser and try again

[brisk tulip]

thanks

[woven stag]

it didn't start though ? it's past 9AM PT

[rustic marlin]

indeed not

[orchid portal]

I tried this and it didn't work for me 😦

Weird. It works in Chrome but not firefox for me

[modest spire]

Can someone direct me to the best channel to ask a question on dall-e having issues within chatgpt since some server issues yesterday? I use it daily to make images for my video stories, at some point yesterday open ai had some issues and chatgpt stopped working. When it started to work again my images seem to generate differently. Now it takes longer, gives updates of each step above picture, and it comes out like a muddy oil painting each time.

My wife doesnt have a paid account but we tried to generate the same image from hers and it worked as mine used to. It said creating image and it appeared in the style I'm used to. But for me, it's doing some new style that I can't use and looks like trash.

Anyone else experiencing this or know how to revert back to how it was producing images before ? Or is there a better channel here to post this question?

[sonic escarp]

let me back in please openai team

[wanton kettle]

Seems like image prompts are stuck at 99% while videos are totally fine

[sonic escarp]

when I was on same thing but I cannot even login anymore

they will fail at 99% too that is why I tried to reset my cookies and connections

BAD idea

[heavy patio]

It's the new image gen model. It's a little different to prompt, it defaults to that style. If you want a photo just ask for a photo instead of image, or ask for a realistic image, or say you want a portrait, or give a reference image of a style you like. If you need help prompting share your prompts and outputs in #prompt-engineering

It listens to prompts much better though so once you get used to the change you will probably like the new model, or at least that was my journey.

[wise matrix]

Hi, hopefully I can get some help here. I want ChatGPT to stop using icons! Yeah I know, weird, but it is really annoying. I added NO icons in responses and suggestions, unless specifically asked for. to my "Custom Instructions", in Memory there are several records of my instructing to strictly follow the "no icons in responses" rule. But the AI keeps ignoring this and continues to use icons.
What can I do to stop this?

[timber shadow]

Hey! Do you mean smiley faces like 😀 ? If so, try changing "icon" to "emoji" in your custom instructions -- "Never use emoji in your responses" etc

[loud prism]

I've got it i think i go wrong website so i start with openai webpage thank god

Now i can use it on pc thanks

[low adder]

posted this in the forum, but anyone know what are considered "requests" when using the real-time API? Is it the number of WebSocket connections? Or is it the number of requests sent over those WebSocket connections?

Cause the Tier 5 limit is 10,000 RPM and I'm trying to figure out if that means if I can have 10,000 concurrent websockets

[wise matrix]

Strangely, I haven't tried that yet... Thanks for the suggestion!

[fading solar]

Hey, is anyone else having trouble turning an animated image into a realistic one? It doesn’t seem to work for me anymore.

[meager folio]

took my money yet didnt get the subsribtion

[knotty bobcat]

Seems to be working again now, out of random, dont want it to happen again tho

[glad cloud]

I keep getting "There was an unexpected error running this prompt" . any one else facing this issue? All wordked well for a few hours, suddnely it keeps giving me this error

It happens on the sora UI, on the chat, It seems to generate images well

[west bobcat]

its happing for me as well, its unable to make any photos

[glad cloud]

Yes. I hope its temorary and will not be long. Its been a few hours now that i can't generate anything

[glad cloud]

now it became "Unable to generate
Service at capacity, please try again later" Instead of unkconw error lol.
OpenAI, Whats going on?

[limber solar]

The studio gilphi style anime image is for paid users right?

[charred parcel]

GPT-4o image gen is for paid users only at the moment

[limber solar]

Any chance it will be available for public in future?

[charred parcel]

maybe a few weeks, although I think it might be longer, the servers are working too hard now as it is

[limber solar]

Yeh so it have chance to become public feature in the future like dalle

[charred parcel]

if you're interested in images or video then a paid sub is the way to go

[limber solar]

Hmmm I can't i will wait

[charred parcel]

free users will get 3 images a day

[limber solar]

So i can use the chatgpt o4 new the image generation?

In free plan?

[charred parcel]

not until they release it, which might be a few weeks

[limber solar]

Already no problem i will wait

[mortal rover]

Not really sure I'm going to find anything useful here. Is everyone else just trying to create images, goofing around, and asking AI to do their homework for them? Did I join in at the wrong time, or do I need to dig deeper? Just looking for inspiration, see what kinds of creativity is going on out there.

[charred parcel]

check out #images-canvas and #images-discussions

[mortal rover]

So just making images? Got it.

[charred parcel]

Australia/New Zealand is waking up, it's midday in California and Paris is close to retiring for the night. This Discord server can become busy, you've just come at a slightly quiet time.

[azure storm]

my account was deactivated this morning i was on pro plan i sent a bunch of emails and i just got back that it was wrongly deactivated but now im on the free plan and the pro plan was just removed from my account

[charred parcel]

did you get an email from OpenAI?

[azure storm]

yes it says thank you for reaching out, we have determined that we incorectly deactivated your account access and now im on the free plan

[charred parcel]

what are you going to do, resubscribe?

[rustic marlin]

for that sort of issue, only the support can help you, you will need to contact them again on https://help.openai.com

[mint sand]

I'm in Nigeria, I tried inputting my card details but all of them were declined. Then I proceeded to fund my dollar card I use for online transactions to use on the site the payment failed but I was debited and refunded instantly. Tried multiple times but same result.

If there's any Nigerian here or anyone that experienced something similar can you tell me how I can go about it.

[slow bear]

In chatgpt is there a way to stop the read aloud option from repeating midway through?

[upbeat mauve]

Greetings!
I'm using the browser-based ChatGPT interface. One of my sessions is locked up. It's been more than 24 hours.
It's stuck typing mid-sentence and the stop button does nothing.
It's not super long, but I'd hate to have to lose it.
Latest Chrome and Brave. I tried logging out, clearing cache, and using a different browser, but it's still stuck. I don't have a phone, so I can't use the phone app.

[sacred fossil]

In Sora, when you click on the remix button next to an image, it allows you to paint on the image with a transparent brush. Does anyone know what that does? It doesn't seem to act as an inpaint because it still changes the generation outside of the inpaint area. Any thoughts?

[winter gust]

is the new image generation working for you guys? it keeps saying that to me idk why

[dense jacinth]

If there a channel related to OpenAI CLIP model?

[tight birch]

Can I change my email address on my account? Seems like no option

[charred parcel]

no you cannot. if you need help then go to https://help.openai.com

[frank light]

#community-help i am a chatgpt plus user...i have uploaded 2 pixar style images.. the one that is drinking tea is latest image model...and one that is hugging is dalle - 3 #image-bot model....now the problem is i want to use dalle 3 image generation in my chat session...but in chat session new image model is been used...and in new chat session i am unable to generate images using Dalle 3...plz resolve this...let us decide which image model we should use

[viral trellis]

hey i am facing so many issues, i have plus account and not able to generate any thing except Ghibli style and that also take so much time and not able to use sora getting error and retry prompt and getting this message

is there a problem with my account, i have good internet speed

please can anyone can help?

[mystic herald]

Sir. My sora AI model is not working. It's been 1 hour but it's still delayed

[viral trellis]

so every one is facing this issue?

[mystic herald]

Bro also encountered the same case?

[viral trellis]

yeah every one is doing this ghibli style and i am fed up with this only too much of use

[dim sage]

true

[mystic herald]

Sir. When will this delay end????

[charred parcel]

I just generated two 720p 20 second videos in 5 minutes

Try not to queue up more than 2 at one time

[dim sage]

after the hype dies down idk...

so boring

[charred parcel]

it's working fine for me, currently generating images on Sora.com and ChatGPT

[dim sage]

because you pro

that pretty much explains them all

has the highest priority

[viral trellis]

u have the pro version

[charred parcel]

yes

[viral trellis]

after hours of waiting i got this

[dark briar]

Hey guyz recently I integrated open ai on a chatbot on webapp but client separated his ways how to detach my api keys and my own trained (may be delete) from open ai playground

[dim sage]

😔

[viral trellis]

if this is happening they should pause this for now and keep to pro users until the server is ready for the load

[deft obsidian]

i didn’t know with plus you have a minimum amount of responses 🫣

i’m about to start midterms and i got demoted until it unlocks responses again in 10 more replies !!

please send help 😞

i def do not cheat or even copy and paste !! i just study with my bestie bot !!! 🤖 we talk thru everything since i am a cyber security student…. coding and chat gpt are game changing when it comes to truly understanding code 🔗

[subtle surge]

yo, this channel active?

[mystic herald]

What is sir's account name. I want to refer to

[rugged wing]

hello ?

who is the admin over here, we need to talk about this

about chatGPT

I want to talk to the admin now !!

[charred parcel]

you can message @burnt galleon if you want to message server staff

[rugged wing]

thanks

[fringe spade]

I'm developing an AI voice chatbot that integrates speech-to-text (STT), a language model (LLM), and text-to-speech (TTS). My STT component relies on OpenAI, but I'm facing an issue: the chat completion feature only functions with OpenAI versions earlier than 1, and I'm currently on version 1.66.
Would downgrading OpenAI allow me to use the chat completion feature while still retaining the other functionalities of the Livekit platform? If so, which version should I downgrade to?
Additionally, the chat completion function isn't working— store:true command doesn't execute as expected. Any suggestions or guidance would be greatly appreciated.

[trail cairn]

do you know what the store:true does?

[smoky plaza]

hello, is there a reason on why it doesn't let me generate any image in sora as it is blocked in preparation still?

[charred parcel]

servers are overloaded, there's a notice on https://status.openai.com

[smoky plaza]

thank you for helping!

[brisk pewter]

Hi . I just wanted to know that whats the image generation limit per day with plus subscription . I am not able to get a definite answer from web . pls help !!

[trail cairn]

because there is no definite answer

[brisk pewter]

yeah , but some plus users might be here , they can just share their experience

atleast approx number , may be

[pale thorn]

The server status..

[trail cairn]

the limit changes every day, or even every hour like now due heavy traffic

[sacred fossil]

Can anyone explain the discrepancy that I'm getting here on sora.com?

[charred parcel]

rate limits due to GPU overloading

[sacred fossil]

Okay, thanks!

[burnt sky]

in chatgpt settings, what is "Show follow up suggestions in chats" ?

[willow ember]

Anybody know how to fix this 😅 Need some help

[brave rivet]

what is the current limit for generating images,i saw they remove the ( unlimited images and video generation) on plans option

has anyone reached limit yet?

[toxic estuary]

hey im trying to log into my openai account but when it says it sent me a opt i never get it in my email can someone help?

[rugged wing]

I’ve been using ChatGPT intensively for the past few weeks to work on a very detailed case involving legal complaints, tax filings, immigration documents, and paystub summaries. I’ve spent countless hours uploading files, organizing timelines, and working step-by-step with GPT to build my final documentation.

But now I’m running into a serious issue: files I upload disappear from the session even just 1 or 2 minutes later. I keep getting messages that “files have expired” or that “the system reset,” even if I just uploaded the file moments ago. This happens repeatedly — not once, but dozens of times in the same day. It didn’t used to happen this often. Earlier, my sessions lasted hours without problems, but now it feels like I can’t make progress at all.

It’s extremely exhausting to re-upload the same ZIP files or PDFs over and over again, especially for someone handling complex, sensitive documents. I’ve created ZIP bundles, backed them up, and even worked around the system’s limits, but the resets continue without warning.

To be clear — I’m not blaming GPT. The assistant has been helpful and responsive. But this system behavior (resetting the session and deleting recent uploads without notice) is causing massive disruption to people doing serious work.

Please OpenAI or community moderators — is there a way to make sessions more stable or allow files to persist longer? Even a 15–30 minute safe window would save users so much frustration.

Thanks for listening. I really hope this gets attention.

[mystic herald]

I'm using the plus plan but the rating waits too long and often crashes when using sora. Will I switch to the pro plan without a rating?

[dim sage]

~90 in chatgpt, ~201 tries in sora (402 images max)

[cerulean blaze]

i cant use voice mode on windows app someone please help me it freezes when i try and use it

[pearl geode]

I have the plus plan but now it's only letting me create 1 image at a time whereas before i could do the max of 2 at a time. Any ideas why

[charred parcel]

they are rate limiting due to load

[pearl geode]

ahhhh thank you, makes sense

[charred parcel]

should be back to normal in a few days

[pearl geode]

cool no worries then

[drifting hawk]

Does it throttle down to 1 generation at a time at some points? I thought based on the text it would be 2 concurrent generations

[charred parcel]

they are rate limiting at the moment due to high load, it should be back to normal in a few days

[gritty anvil]

Is their a current reason why almost nothing works imange wise? The more detailed I got on a prompt (using a prompt engineer) to get what I wanted every GPT I've ever used crashed saying the prompt was basically too complicated

[brave rivet]

does it mean we can generate unlimited images ?

[charred parcel]

not at the moment, you will be rate limited

[rugged wing]

Sever down?

[rotund bison]

Hey, I’m experiencing a persistent issue with the “Work with” feature (Direct Edit) in the ChatGPT desktop app while using O1 Pro mode. It keeps returning “Unable to generate edits.”
• Regular O1 works fine.
• This only breaks in O1 Pro.
• No VPN is used on the device.
• Already tried reinstalling, logging out/in, rebooting, etc.
• Using Mac mini M4 Pro, macOS Sequoia 15.1.1

This seems like a bug in Pro mode, not a user-side issue. Any updates or workarounds?

[rotund bison]

Same problem

[ancient herald]

Welcome to my world. No one from openai has ever addressed nor talked about this issue so far.

[charred parcel]

Work with hasn't worked with o1 Pro for a while

[ancient herald]

"it never" worked for me

[rotund bison]

Support has been ignoring me for the second day already.

[charred parcel]

Support is very busy, they're not ignoring anyone, just trying to work through the load

[ancient herald]

how do you know?

[charred parcel]

because I had an email response from OpenAI a couple of hours ago

[ancient herald]

I didn't email them, I opened a bug-report here

[charred parcel]

Make a direct report at https://help.openai.com

[ancient herald]

how?

[charred parcel]

make sure you're signed in and use the messages button on the chat bubble

[ancient herald]

where exactly can I sign in?

[charred parcel]

if you're not already signed in go to https://platform.openai.com and sign in

[ancient herald]

I think I found it. Have to go before to the messages tab

[rotund bison]

It’s already been 18 days since the issue first appeared for Steph, and two days for me. I’ve already contacted support directly, but I’m being ignored again.

[ancient herald]

I even replied to X to Desktop app developer and he was replying to other people but didn't reply to my issue report

[charred parcel]

you're not being ignored, support can sometimes take a week to reply

It's really hard to scale a support desk when your userbase is 400 million per week

[hot crescent]

anyone know how to cancel the subscription ?

[trail cairn]

manage payment

[rotund bison]

Yeah, that all sounds great, but in such critical cases — especially considering the price of the monthly subscription for the GPT-Pro model — support should be putting in much more effort. Especially in such a highly competitive time in the development of AI models.

[hot crescent]

i can't do anything there, i can't remove my pay,mentcard or anything, since it says it has an active plan on it.

[ancient herald]

18 days no response is not great considering I'm paying 200$ for 30 days of use...

[charred parcel]

I am a Pro subscriber, you just need to be patient. If that feature is the only reason you have Pro then unsubscribe because I don't know how quickly they're going to be able to fix it

[ancient herald]

I'm surprised you're the first other person except me, reporting on this bug to be honest

[rotund bison]

Perhaps it is worth appreciating individual developers who make up the majority of the actively developing OpenAI community, in the current reality I increasingly see a preponderance of interest in favor of Sora with the lack of normal support for chat GPT

[charred parcel]

there are individual teams for each feature nobody is taking people off teams in favour of other ones

[agile meadow]

?

[stuck terrace]

is anyone also getting this error when uploading images to chatgpt

[stark haven]

??

I've paid for the services, but there's still some problem.

[rotund bison]

Could you please notify the technical support team about this fundamental issue (since the tools feature is exactly that), and at least provide some logs to help understand what’s going wrong? I understand that support is busy, but if this issue isn’t widespread, maybe there’s something I can do on my side to fix it.

[charred parcel]

no I can't, I don't work for OpenAI and I am not server staff

[rotund bison]

Secret agent 😎

[nova jasper]

anyone know what's causing this? am i the only one?

it was fine the day before yesterday

yesterday whole day was like this

[open topaz]

Where you get these settings

[nova jasper]

sora

[open topaz]

sora.com ?

[gritty anvil]

Honestly it's not that hard to scale, OpenAi is the single most biggest disruptor since Uber/Lyft & AirBnb, they have raised $17.9 BILLION, they literally can take the some of the 900,000 to scale up the basic support staff for issues like this.

Like basic CS to say "have you gone to this option" "sir/ma'am I'm sorry but the GPT must have found it suggestive and didn't complete the query" it's not taking any creative job no time soon..lol.

Like they have the money to train AND scale, they choose not to or it's not even a consideration for them

[zealous zealot]

hey hi, are there any guidelines how to prompt sucessful for sora? i used some guidelines, asked chatgpt but the video i created always is messed up and sora completly changes the setting during the video

[west bobcat]

Has there been an announcement saying why we can only do 1 instead of two for plus. the subscription plan info still says 2 and hasn't been changed to one.

[charred parcel]

there are rate limits in place

[west bobcat]

so is it for me or everyone?

[charred parcel]

everyone, Sam tweeted about temporary rate limits the other day

[brisk pewter]

are images getting generated now . my dalle says "high demand " and not generating at all. plus subscription

[junior canopy]

The search and filter feature on the SORA website does not work properly. You can search for a term but when applying filters nothing happens even after reloading the page with the filters.

[west bobcat]

in the image remix, what's the point of selecting an area to edit, if it edits the entire image even when you don't highlight it

[placid bay]

I am so confused right now. My subscription was essentially terminated because of some "Suspicious activity." I did everything it asked me to do and it still doesn't work. What is going on?

this is actually making me mad

[solid cairn]

Hello, where can I manage my subscriptions? I have been looking for 10minutes and I can't find it, openAI their sub domains are like Dungeon

[green wolf]

Account creation curently unavailable on Sora - will it be fixed?

[trail cairn]

its likely disabled until they get more capacity

[green wolf]

thank you

[frank vortex]

is there a place in here to learn all best practices for image creation in Sora and/or best practices for interacting with chatGPT? I'm talking optimizing my inputs for it to understand

[severe heath]

think that qualifies as 'prompt engineering' in general, there are some prompt tips/engineering channels on the sidebar, but you can also ask gpt itself for help in optimizing

[finite sierra]

One message removed from a suspended account.

[gritty anvil]

I've been getting that since Friday night, at this rate, Deep Seek will overtake them & they are using dated chips

[compact sage]

Guys, don’t get your hopes up too much with this — not while it’s still working like this. For now, just treat it as something fun to mess around with when you’ve got time to kill. If a half-baked product gets released too early, it’s better not to expect too much from it… at least for now.

[gritty anvil]

Yeah but this is such a bad business practice for OpenAi, like youbhave Deep Seek right at its door and you can't do full capacity?

[compact sage]

Why would they even care about us? We're just small-time paying users

[gritty anvil]

You do have a good point lol. I'd be pissed if I was a $200 paying customer

[compact sage]

$200 is probably just pocket change to them too. Maybe it gets a tiny bit of attention, but to them, we’re still just ants in the grand scheme of things

🤣 🤣🤣🤣🤣🤣

[gritty anvil]

Lol, I have to play nice now, server mod put me in time out for 2 mins..🤣🤣

[finite sierra]

One message removed from a suspended account.

[novel pebble]

My chatgpt keeps telling me 4o image generation isn't released yet.
It generates everything using DALLE. I'm on the $20 USD plan, should I not have access to the newest image generation?

[rotund bridge]

hey guys i just upgraded to chatgpt pro, but now for some reason the reason option is gone for my paid account, which is still there on my free account, why?

[gritty anvil]

I've just stop trying for now, rather then get frustrated

[frank vortex]

Can anyone link me to a guide of how to create the perfect face anchor of myself to make AI imagery that resembles me?

[short patio]

its in the model selector o3 mini o3 mini high and o1 pro mode and o1 are all reasoning models (thinking)

[rotund bridge]

thanks it would seem so lmao, just wished they mades a heads up👍

[short patio]

yea, np

[desert fulcrum]

wthehell is this for a paid user lol

[west bobcat]

i guess you aren't special enough /joking

[barren holly]

Hi I’m new and this has probably been asked a million times so sorry in advance … when does iPad version get updated to this new image model?

[charred parcel]

what iPad version?

[barren holly]

Oh sorry I just realised I look at it on the web but on the iPad lol 😂 told you I was new

[charred parcel]

are you a free or paid user?

[barren holly]

Paid…

[charred parcel]

then the new image model is available to you

[barren holly]

Plus? I think? How do you use it?

[charred parcel]

just ask ChatGPT GPT-4o to create an image eg:create an image of a cute puppy

[barren holly]

It says it’s still using dall e?

[charred parcel]

then you must be a free user, DALL-E has been retired

[barren holly]

I’m on plus?

£20 a month?

[charred parcel]

don't ask it which image model it uses, it doesn't know. Just do what I said create an image of a cute puppy

[barren holly]

Oh Kk

[hot pulsar]

How do I add Chat GPT to a private server? Is that possible?

[charred parcel]

no, it's SaaS served from OpenAI

[hot pulsar]

SaaS?

[charred parcel]

Software as a Service

[hot pulsar]

oh...

[charred parcel]

anyway you can't run ChatGPT locally

[barren holly]

It works :) Thanks

[charred parcel]

that's the new model for sure

[finite tiger]

chatgpt say " You have reached the maximum number of prompts you can send per day. " but from what i know limit is 80 messager per 3 hours 🤔

[dark hawk]

Hey guys

I'd like to buy the Plus subscription, but I was wondering about the limit on the new image generation. Could anyone tell me?

[trail cairn]

changes every day and sometimes every hour so dont buy anything just for that

[dark hawk]

Oh, alright, but it should give plenty of images, right?

I'm looking for 20 without a limit...at least

[trail cairn]

not plenty

[dark hawk]

Meh, some will do

[trail cairn]

servers are overloaded

[dark hawk]

It seems like it

I'll give it a try though

Thanks @trail cairn

[onyx kraken]

I needed to uninstall, reinstall my ChatGPT OpenAI app, and then it was using the new image model for me

[onyx kraken]

What kind of prompts?

It's possible that some prompts, like image gen or other specific tools could have a lower limit - but we can still send the normal messages to 4o or the other models.

[charred parcel]

a few people have said they haven't been getting 80 messages per 3 hours, maybe OAI are rate limiting all of GPT-4o

[onyx kraken]

Ahh, thanks for the heads up! Wasn't aware of the comments. Remembered the early Dall-E-3 days, when we had so many images per time periods

[potent magnet]

Hello, im having an issue with my subscription, i took a gpt+ subscription, i got the confirmation that the sub is active but i can't use gpt+ it says i don't have the right but when i try to upgrade again it says i already have one. I payed it with Apple Pay on my phone. Thank in advance for any help

[onyx kraken]

Maybe could be a cookies or similar issue, maybe try restarting the app/reloading the webpage.

You need tech support if there's more to it, bot bottom right corner to start the process.

https://help.openai.com/en/articles/6614161-how-can-i-contact-support

[finite tiger]

i'm on Plus subscription and i use gpt 4.o lates version

[winter mantle]

hello i have a question, how do you use whisper word timestamps

[onyx kraken]

What model are you using? On plus, you should be able to at least use the 4o-mini.

I haven't noticed anyone report that error before!

I'm a community member like yourself, but I do try to notice what I can that's shared on the Discord.

I checked the helpfiles.

I don't see mention of that either.

I'd definitely ask tech support, you can share the screenshot through there too.

You can also #1070006915414900886 if you'd like to on the Discord, otherwise the website tech support's fully private between you and OAI.

The discord's bug report, if other users are affected similarly, they might add to your report with their own details.

[finite tiger]

like i write in my message " 4.O late version "

[graceful locust]

Can someone provide me.a free ai tool to create Ghibli images

[thorn falcon]

I am trying to find people who want apps that utilize ai in specific ways and would actually pay to use them, where do you figure I should find such people to discuss with them and start building for them and with them?

[charred parcel]

become known for your work, leverage LinkedIn

[thorn falcon]

hmmm, known for my work, that makes sense

I can make ai powered apps and websites pretty well pretty quickly, how do you suggest I become known for my work? do i make apps out of my own head and spam post them on linkedin?

Do I post about my knowledge of Ai there?

Also im not trying to do commissions but rather build apps that are purchase based or sub based, I just dont have the audience nor the marketting ability to do so, quite a pickle

[charred parcel]

this is beyond the scope of the OpenAI server. You just need to build a following. Release something for free and Open Source it, put it on Github then show it off

[thorn falcon]

I see, that makes sense

[onyx kraken]

Our community has a rule against various things related to self promotion, #server-rules , is why it's beyond the scope.

However! ChatGPT itself, in your own chats with it, might be an excellent resource for you to personally explore for ideas how to become more known for your work!

And asking your opening question to the model also is likely a good idea to explore, both with and without search.

Good luck!

[thorn falcon]

thanks, im not trying to self promote or anything im trying to figure things out i havent made a single dollar online yet, sorry if it sounded like that

[onyx kraken]

We have to be careful because we need to follow the rules too! And can't encourage others to possibly break them.

But I totally would ask the model, research, discuss more with the model, as well as any other ideas you have for the right places to discuss how to succeed.

[thorn falcon]

yes thanks, I'll do just that

[unique pewter]

I’m currently doing a university assignment which requires me to use ai sora. Unfortunately, it says new accounts can’t create videos and it’s due on Friday. When will I be able to create videos? 😅

[charred parcel]

we don't know when new signups will be allowed

[unique pewter]

Do u reckon I will be able to by Friday?

[charred parcel]

depends on whether OpenAI can get the load issues resolved

[unique pewter]

I see

[uneven vigil]

Rn every of my image generations in Sora is failing. Does anyone have an idea why this might be?

[charred parcel]

share one of your prompts

[uneven vigil]

This one failed: A black and white street art stencil on a wall in the style of Banksy depicting Lenin as a humanoid half-robot

This worked: A hyperrealistic black and white street art stencil depicting Lenin as a humanoid half-robot, silkscreen print

[charred parcel]

the first one is a content policy violation

[uneven vigil]

Why? Because of Banksy? Street art stencil on a wall?

Are you sure btw, I had other content policy violations that had been clearly flagged as such.

[charred parcel]

Banksy I would say

[uneven vigil]

@charred parcel Thanks. That was it. It's a pity that the filter is so restrictive.

[charred parcel]

my advice to everyone is steer clear of IP

[dry bone]

hello, I've been having some issues with long conversation and analysis. The problem is that the browser usually freezes forcing me to ither refresh or open a new tab to see the answer. Is this a problem of my computer? browser? or ...? I'm not sure where to start looking for a way to fix this.

[charred parcel]

start a new chat

[dry bone]

It's hard for me to do that, the analysis i've been doing is extensive and exhaustive building step by step (the results and analysis is built in the previous answers). if i start a new chat i'll lose all the previous work :/

also don't think a "make a summary" and start a new chat would cut it for the level at which i'm doing this

[charred parcel]

are you on Plus or Pro?

[dry bone]

plus

[lyric ruin]

Are the servers down at the moment?

[charred parcel]

ok ChatGPT can only remember the last 32k of the chat, just ask ChatGPT to create a summary then copy that to a new chat and continue

[dry bone]

i see... i didn't know that.

thanks for the support I'll try that

[upper flame]

im guess chat gpt is down? been doing this for the past 45 min

[fallen sequoia]

What does ‘this image has been canceled’ mean?

[fallen sequoia]

I wish there was a notification on the OpenAI Sora page that there are problems like this at the moment.

[young roost]

Hi , quick question , if i take Pro Plan , should i be able to use image generation without limits and faster as well?
or it's effect only on Video ?

[charred parcel]

yes

[modest fiber]

Guys could you explain me implementation of API, like how devs use that and what are those api keys
I heard about mcp - and that is somehow leads to some ai chatbot controls software
and if it is protocol is it just like - method ?

[fringe spade]

Does anyone knows how to replace stt-llm-tts pipeline wih opeani realtine api in the following code.I have taken this code from restaurant.py.
code link->agents/examples/full_examples/restaurant_agent/restaurant_agent.py at main · livekit/agents

[merry mist]

I have plus plan. And I had the possibility to generate 2 images on queue. Why from about 1 hour it only makes 1 photo on queue and it's written that I have to update my plan. If i have plus and I can generate 2 for queue

[charred parcel]

Image generation has been rate limited due to server load

[merry mist]

Ok thanks. And for videos when they will unlock it again?

[charred parcel]

what happens when you try to create video?

[merry mist]

It says new members cannot generate videos for now. This Is cause there are too much people using videos

[charred parcel]

it's because servers have been overloaded since the new image model was released, they are trying to alleviate overloading GPU capacity

[fallow finch]

Question, is it expected to have video disabled? I just upgraded my plan to pro and I still see the video disabled apparently becuase my account is new. But I've had this account already with Plus.

[gusty vortex]

Thanks for the info, i have the same issue, with images and videos; and the videos are taking like 15-30 minutos to generate a 5second video

[hot brook]

I never thought I’d run into my Deep Research allowance limit on Pro but somehow I did and I have to wait until the 8th of April. It quickly became my favourite feature. Just what exactly is the allowance on Pro, so I don’t run out asking it for another research paper on why people think missing a train is the worst thing in the world?

Jesus, I did more than 120?

[tough axle]

Is there a reason the deep research credits I had set to refresh on April 10th got pushed to the end of the month?

[stuck terrace]

unable to upload images on chatgpt plus

also for some reason looks like this

[charred parcel]

reach out to support at https://help.openai.com

[dark hawk]

Servers down. Try later on...it's the same for me

[lilac thunder]

This is the worse time for servers to be down.

[mellow marten]

Heyy, I was using chat gpt and suddenly collapsed (literally). Now I have a chat that I can't delete, neither access. This happened me 1 week ago also, and I can't find a solution to delete the chats (currently 2). Anyone with the same problem?

[raven vapor]

I need help from a mod pls

[timber shadow]

Hey! Send a DM to this account: @burnt galleon and select the option to reach out to server mods

[vale prawn]

man need help with assistants; i keep getting a "message': "The v1 Assistants API has been deprecated. Please try again by setting the header 'OpenAI-Beta: assistants=v2" but i do have the headers set to v2

nvm all set 😄

[spring kestrel]

Subject: Urgent Follow-up – Ongoing Image Generation Tool Issue Despite ChatGPT Plus Subscription

Hello OpenAI Support Team,

I am a ChatGPT Plus subscriber, and I previously submitted a complaint regarding the malfunction of the Image Generation tool. Unfortunately, I have yet to receive any updates, despite a long wait and my full compliance with the usage policies.

Since this tool is a key feature of my paid subscription, its prolonged downtime without clear communication or a defined resolution timeline has been highly frustrating. I subscribed to this plan expecting a faster and more efficient service.

Therefore, I kindly request an official response and urgent clarification regarding the following:

• The reason for the ongoing issue and why no internal notice has been provided.
• An estimated timeframe for resolving the problem and restoring the tool.
• Possible compensation for the period during which the service has been unavailable, ensuring I receive the full value of my subscription.

I hope this matter will be taken seriously, and I look forward to your prompt response.

Best regards,
Abdullah

[charred parcel]

you should send that to support at https://help.openai.com as nobody here can help with your account

[spring kestrel]

No one answer there

[charred parcel]

support can be slow, be patient or resubmit the request

[green magnet]

is the image generation limited to 1 prompt at a time to plus users? it used to be 2

[charred parcel]

yes, that's the current rate limit

[green magnet]

server overload?

[charred parcel]

yes, they are rate limiting to keep the service operational

[green magnet]

ok. Any ideia when it will be back to normal?

[charred parcel]

I don't know, sorry.

[green magnet]

thx bob

[spring kestrel]

Okay i will , thanks

[still fox]

Hi guys. Are you experiencing trouble making videos in sora? It keeps saying disabled to new users, but I’m not new? Have a plus plan and have had that for a while now

[charred parcel]

did you only just sign up at Sora.com though?

[still fox]

Ah yeah I just did that

[charred parcel]

that's why you need to wait

[still fox]

I see, ty for responding mate

[fringe spade]

can the openai response api be used asynchronously?
facing error->2025-04-02 00:29:35,802 - ERROR voice-agent - Error enhancing transcript: object Response can't be used in 'await' expressi

[humble wagon]

does anybody know if its possible to upgrade from a chat gpt teams account to a pro account to access full sora settings?

[versed kraken]

Can you please tell me what I should do? I bought an account specifically on Sora to realize a project on video animation. But I get the message that on new accounts video does not work. I paid $ 200+ for it and I'm on deadline for the project. It turns out all the money is wasted(((( What should I do?

[past spade]

Hey All, I'm new here, I've been playing with AI for a while and use it to automate some of my job, but I've just started playing around with Sora.

I'm learning how to craft my prompts to return images that are at least close to what I'm trying to accomplish.

I'm running into an issue with getting Sora to accurately produce the images portraying figures in the positions I want, from the angles I want to see them.

I've tried tons of different ways to word things or verbally describe the way in which I want the images to look, but I keep ending up with images that force things into certain positions.

. . . for example, I'm trying to make a picture of someone riding a dragon, standing on it's back, rather than riding it like a mount. . . but no matter what I say, it's trying to make the person ride the dragon like a mount.

[junior marlin]

Hi there 👋

What version of GPT is currently used in Custom GPT? Is image generation inside Custom GPT’s with the new image model compatible yet? I only seem to have access to Dalle in the custom GPT’s at the moment. Thank you in advance for an info you can help with 🙂

[onyx kraken]

There's a number of ways that probably would work. This is the one I'd try first, there absolutely are others.

So, focus on describing the unique details that require standing to happen, and describe the position of the feet, the stance, the weight distribution, where the arms are, the expression on the face, as well as the dragon, what it's doing, and the rest of the scene.  Photo quality.```

[onyx kraken]

From what I understand in reading the helps.

You would need to cancel the entire team (if you are the owner) or remove yourself from the team.

Once that is done, then you subscribe to pro.

"Can I access ChatGPT Pro features from my ChatGPT Team account?

No, you cannot access ChatGPT Pro features from your ChatGPT Team account.

To access ChatGPT Pro features you can create a new account under a new email to sign up for ChatGPT Pro.
​
Users with ChatGPT Team accounts can cancel their account and update their account to ChatGPT Pro once their subscription ends."

https://help.openai.com/en/articles/9793128-what-is-chatgpt-pro

[timid hamlet]

hey people.

Urgent help needed! This is a crisis. My OpenAI account was abruptly deactivated with zero explanation. I've contacted [e-mail] several times but received no response. I'm in a nightmare scenario – years of work, significant financial investment, and my entire business built on ChatGPT are now inaccessible. This is a devastating loss. Has anyone experienced this arbitrary account termination? How can I escalate this or pursue legal action? This is catastrophic. How can anyone trust OpenAI for business development if they can pull the plug like this?

[onyx kraken]

Sorry to hear you're having this situation.

I and everyone here are community members like yourself, the extremely few here who work for OpenAI have gold names, #server-staff.

Yes, people do come to the discord sometimes and discuss having had their account suspended.

Trustandsafety's a new email available, previously people needed to reach out through https://help.openai.com/en/articles/6614161-how-can-i-contact-support as the only way possible to ask for help, and I believe that option's available still as well to discuss and seek assistance.

It has typically taken minutes to weeks for a response from OpenAI help tech support. That's the range I personally have seen in my own requests, and also noticed others talk about with their own requests.

I recommend great patience, difficult as that may be, because I understand that OpenAI has more than 400 million weekly ChatGPT users alone (and they have other models too) and they recently reported over 1 million new accounts joining in a single minute.

So.

They likely have unprecedented levels of interactions and requests for help to sort through and reach back to each person needing help.

I don't know of any option to speed the process. I would try to remain calm and helpful, so when they can help me, help is possible.

For any urgent needs I had right now, understanding that an instant resolution is likely simply not possible, I would use anything else I could to meet the needs during whatever interim.

I'd treat it more like a power outage or the internet being out than something I can personally control.

Good luck, I really hope you hear back very soon!

[clear pilot]

Hey there looking for some help with a couple of hopefully quick fixes here. I'm subbed to the plus 20 buck a month plan and the image generator is telling me I can only have 1 image at a time and also I can't generate videos

How do I fix these issues. Thanks for the help

[timid hamlet]

Thank you for your kind words and for trying to help. I understand you're a community member, and I appreciate your suggestions. However, 'patience' is a luxury I can't afford right now. While OpenAI sorts through their 'unprecedented levels of interactions,' who's feeding my son? This isn't just a technical glitch; it's a complete disruption of my livelihood.

I found others experiencing similar issues. That's precisely why 'waiting' isn't an option. This isn't just about my account; it's about the countless developers who've built their businesses on OpenAI, only to have them arbitrarily taken away. I'm starting to wonder if raising media attention on this systemic issue could help ensure other developers are aware of the risks before investing their time and resources. Because right now, the lack of transparency and support is simply unacceptable.

I'm not looking to be difficult, but I need a resolution, and I need it now. And if OpenAI can't provide that, then perhaps a wider audience needs to be made aware of the consequences.

Sorry to use your well intended answer for further information. It was also intended for other potential readers who may be in the same situation, or considering working with the platform.

[onyx kraken]

We're all community members. I appreciate you discussing what's going on, one of the things I attempt is to track what folks share and try to mix solutions to problems for the future.

I'm sorry I am not aware of any way to hurry attention to you. I didn't cause your situation, and I can't resolve it, I'm a community member like yourself.

I do wish you the best.

We can all speak as community members here, inside #server-rules. Good luck on a desired and rapid solution.

[tacit pike]

Hi Folks, I am looking for a summarization model especially to summarize the dialogues / conversations between 2 persons. The model should work on a CPU. I have tried bart-conv, flan-t5, dialogled. But the accuracy of these models are not that great.

[fair quiver]

Hey, i'm using ChatGPT Plus for almost a month now. Since i started paying for it, it seems like getting slower and slower. I have one chat with a pdf attached to it that is 600 sites long. don't know if that is the problem, but all other chats are also slow or not even responding. Today i saw that he didn't save the answers only my question, so have like 6 qustions and no answers. I already tried different versions but its the same, same with other browser or PC, places or smartphone. Sometimes he also says that i have the max. lenght of the chat reached. Any Ideas?

[dapper latch]

i think server overload... good luck for u im with you i have been plus and beta user since 2023 february my main goals for studying university MA and in the future PhD... but i have noticed that what u wrote🥺🫩😢🫂 we've had a terrible time working together since he suddenly had a crowd around him... and even if something meaningful would draw a crowd... but no... so it's really nokomment....😓

[wicked oasis]

i keep getting limit after one messege
also other chats are same...across platforms..cant use it anymore any help?

[charred parcel]

what plan are you on?

[wicked oasis]

Free?
But this is after just one messege?

And it's happening across chats and platforms.

[charred parcel]

Free plan is heavily rate limited right now, either upgrade or start a new chat

[wicked oasis]

So it's a bug right?
And it's happening to new chats as well?

[charred parcel]

have you tried starting a new chat?

[wicked oasis]

I did... This was a new chat.. there was a red error... After two messeges and when I clicked retry I got this.

[charred parcel]

you might need to upgrade then, as I said Free users are being dynamically rate limited at the moment, maybe this is the extreme of rate limiting

[wicked oasis]

You can see other messege line above it...
First I said is this working...
I got yes from chat gpt followed by red line error of some sort I typed my messege and clicked retry...I got limit

[charred parcel]

what was the original error message?

[wicked oasis]

Uhhh... Idk ...it was like .... alphanumeric gibberish...

[urban cloud]

Currently using Pro subscription, considering getting an Enterprise tier instead for a larger memory among other features, yet am barely making full use of Pro features yet ah

[charred parcel]

you can't get an enterprise subscription unless you speak to Sales and have lots of employees

[wicked oasis]

I cannot buy it coz it very expensive in my currency...plus i can't use it's advanced features anyways... I will be fine even if I get old chatgpt 3 something

[urban cloud]

How many employees does it need, I am a company director.

[charred parcel]

somewhere around 100 seats minimum

[urban cloud]

What may that cost?

[charred parcel]

Sorry, I can't suggest a solution. Free plan is being rate limited at the moment.

[charred parcel]

thousands per month

[wicked oasis]

Will it get restored to allow free people some use or is it here to stay?

[charred parcel]

the rate limit will be lifted when server load becomes a lot lower

[junior canopy]

The sora website is just BLANK when i try load it and ChatGPT just keeps saying Something went wrong while generating the response. If this issue persists please contact us through our help center at help.openai.com.

[jolly kelp]

uhh is anyone having issues with Chatgpt right now i'm getting this Something went wrong while generating the response. If this issue persists please contact us through our help center at help.openai.com.

[fierce pecan]

Help me! I paid for the gpt chat plus subscription at 5 am, but so far I don't have access to the features

[earnest verge]

Yeah we all have issues.

[fierce pecan]

There is no way to solve it?

[earnest verge]

Probably the servers, maybe overloaded or something, just have patience and wait for a solution from the OpenAI team.

[fierce pecan]

Has this happened before?

[earnest verge]

I am not sure, I started my subscription yesterday myself, and we both joined this server today, I guess a lot of people will join this server now, when this problem occured.

[charred parcel]

yes, about 18 months ago OpenAI started waitlisting new ChatGPT users because of load

[gloomy snow]

Can anybody explain to me why OpenAi is so 'off' recently; not able to do follow-on prompts; I normally use a 5 whys, strategy going deep. I'm limited today to one prompt before I get this error, and then need to open a new context window to restart; one prompt per window sucks. I'm an AI workshop trainer, and the system just times out, won't accept more than one prompt at a time; I'm having to pivot to gemini, claude, and copilot to finish; what a disaster for marketing and first impressions for OpenAi. How can I improve the online experience to have the necessary access to get prompt text window consistency

[verbal crag]

#status

[minor geyser]

is anyone else finding the newer chatgpt models to be horribly innacruate? it's getting to the point that 95% of code is wrong. there's made up class methods, object attributes and sometimes it just flat out says something isn't possible until i provide documentation as proof.

this is mostly with o3-mini-high which is supposed to be "great at coding and logic".

[past spade]

Ohhh, sorry for the long delayed reply, That looks like it might fix my problem, I'll give it a shot! Thanks!!

[tender moth]

purchased Pro for sora video only to find out it is disabled for new accounts!

where is this explained anywhere? Pro is advertised as including video. Has any new user gotten it enabled?

[onyx kraken]

So sorry this is happening for you and other new subscribers!

We've seen similar things happen with previous new opportunities, where for a time (and it's been a month once) couldn't sign up for or access specific new features or accounts while access support was being updated.

We are informed about this delay on Sora access through the OpenAI helps about Sora:

https://help.openai.com/en/articles/10245774-sora-billing-faq

https://help.openai.com/en/articles/9957612-generating-videos-on-sora

https://help.openai.com/en/articles/10877094-creating-images-on-sora

Among other helpfiles.

The information may be stated elsewhere too, I'm a community member like yourself and the helps and announcements are where I find most of the information myself.

And if you click on the help center link on OpenAI's sora's page, it leads to a help announcing the delay for new accounts.

https://openai.com/sora/

[trail thistle]

having trouble with Sora prompt writing - couldn't put it in simpler terms, still turns out to be a violation. Any advice?

[still fox]

So when do we get access lol

[onyx kraken]

Great question! I am sure as soon as that's known, we get told!

When I've seen similar bottleneck-like situations before - last big one I can think of was after something was added to plus, and like instantly everyone wanted to sub or sub back to plus - and they shut down new plus memberships.

For like 5 weeks.

Which was really difficult for the people who wanted to get a plus membership for other reasons than this new feature or set of features, they just wanted more than free membership offered and couldn't get that until the more GPUs or whatever else needed were ready.

I guess this time maybe folks can go Pro - so they can access what can be supported inside Pro - which is a lot - and just what can't be supported yet isn't.

I run into image gen limits recently, so I mostly stopped making any images, I'll leave mine for those who need it more, except the few ideas I really want to test asap.

But there's no benefit for anyone to restricting access more than needed. I'm sure as soon as we can all be supported, access increases!

[thorny crown]

How come 4o wrote 2 responses . 1 with reasoning 1 without????

[onyx kraken]

Two answers to 1 input?

That is usually preference testing where they have two different models answer you - and ask you to pick the better answer.

Did you get a choice like that, or something else happening?

[burnt marsh]

gpt 4o vs gemini 2.0 flash vs notebooklm for document analysis? Mainly care about which one has the least hallucinations

Will be using to create briefing documents on college readings so that I can read them and extrapolate discussion points for a seminar style class

[fierce pecan]

Guys, I paid for gpt chat and until now my access has not been released, I paid at 5 am

How do I request a refund?

[onyx kraken]

https://help.openai.com/en/articles/7232895-how-do-i-request-a-refund-for-chatgpt-plus-or-chatgpt-pro

Sorry you had trouble. I hope you paid through OpenAI and not anything else that tricked you.

[fierce pecan]

I paid for Google Play, I always paid for Google Play last year and I never had this problem

[onyx kraken]

Google play... which app inside Google play?

[fierce pecan]

Play store

[onyx kraken]

Some apps pretend to be OpenAI ChatGPT. Inside play store you pick which one you choose to pay.

Only OpenAI ChatGPT is the one.

[fierce pecan]

I always paid by signing up within the gpt chat and it redirects me to the play store, I never had this problem, the only difference this time is that I made the payment via Pix

It also appears that I have already made the subscription plan

And I saw a report on the forum from a user with the same problem who paid with a debit card.

[onyx kraken]

OKay. So you paid before and it was fine.

This time you pay again, and it doesn't seem to work?

I translate the image, it seems you are being told, "It was not possible to change your subscription plan. Manage subscriptions"

I haven't heard about this before - and I'm a community member like you, we all are here (the few who work for OpenAI have gold names).

But you can reach out to tech support about billing issues here: https://help.openai.com/en/articles/7231930-how-can-i-contact-support-about-billing

[fierce pecan]

I requested a refund a few minutes ago, almost an hour ago, but I haven't heard back yet.

[onyx kraken]

OpenAI has over 400 million weekly different users.

They recently said that over 1 million new accounts joined up in a single minute.

Even before those levels of people to help, it was common for tech support requests to take anywhere from a few minutes to a few weeks to be answered.

It's also currently after 9 PM where the headquarters are, in California, USA.

Help comes, and the tech support on the help page is the correct way to get into the queue. But there's nothing any of us can do to speed the response.

[thorny crown]

what even happened

[thorny crown]

no it was legit 2 answers no preference testing. then i switched out of chat and back and the non reasoning one dissapeared. weird how 4o was reasoning...

and for the past month for the preference testing. they have wrote the exact same response for both

[onyx kraken]

How unusual! Except not if you're giving inputs with high probability 1 right output (like 1+1=?)

If you would like, this form is a decent way to share detailed info back to OpenAI about weird or bad (or even great) model behavior.

https://openai.com/form/chat-model-feedback/

Can give them the chat link, they can follow it to check it out as well (doesn't need to be a share)

[tired summit]

Do we have live support? I think my account is already been a year, but sora said my video generation is disabled for new accounts

[charred parcel]

that is correct, new Sora accounts need to wait

[strong kraken]

Hey Community! I bought a Pro Account a couple of days ago - to create videos!, there was no informattion in this process that video creation is not allowed for new users. After I bought the package, I got this information when I tried to create videos. that really unfair process, 200 dollar for not to allowed to can start for what i paid makes me really angry & sad.

I sent an email to the sora team, but also didnt get any feedback back!

[charred parcel]

everyone is being treated the same, new accounts need to wait to generate videos