chatgpt-discussions

[inland heron]

in what ways could GPT4 be superior to GPT4o? It only has about a month more data, and the intelligence should be about the same

[light crane]

It is the same and better in some ways. Some people report much better performance, some report the same or lesser. I'm not really sure why this is the case

[torn cloud]

Can someone help me confirm where the 'memory' feature is available to UK subscribers and how one might use the 'video call' feature of GPT 4o ?

[dark pier]

Has anyone tested the limits of gpt-4 in the free version? I don't know how much the message limit is, but now I can only use it again tomorrow at 5:01, I don't know if it's in the morning or afternoon...

[granite harness]

4o has some strange behaviors that inhibit it from performing certain tasks. When using it as an editor/writing assistant, it will often (stubbornly and repeatedly despite attempts to course-correct) simply echo a rough draft rather than provide any advice or rewriting any of it to improve grammar or expand the text in any way. GPT 4 does this just fine with minimal prompting. Perhaps a bug they can work out, but as of yet, they have not.

Quoting another user who had a similar issue with a different workflow:

After getting the answer I tried to provide feedback and ask gpt 4o on how to find a middle ground between paragraph and long spaced out lists. 

What happens now is   gpt4o is just rephrasing the second answer about code no matter what I say.  It is like it doesn't understand at all any subsequent promt.
It took around 6 prompts asking to stop talking about loops until it acknowledged it, and only acknowledged it, didnt answer actual question.  There is something really off with this 4o.```

[silk blade]

try sharing a link to the conversation to 4o in a new chat, and ask it to take over where the previous instance went wrong? used to work the first time browse was new

maybe warn it to not follow any loops

[keen meadow]

In my experience - and I think for almost everyone - the quality is the same. Why not actually provide an example of how it has been poor for you.

[light crane]

Gotta be user error

Or a glitch of sorts

[opaque oxide]

bug with ChatGPT 3,5 calling functions on a free user chat(no subscription). The backend also never accepts the requests.

https://chat.openai.com/share/d39e7836-2f52-43d1-bda5-c0a9189d7635

How do I report this to OpenAI

this isn’t a jailbreak, and is annoying to me trying to implement code interpreter myself.

[maiden fractal]

3.5 isn't good for code, also the preprompt like that isn't a good option .

Also 3.5 free have a small limit of context

[tight zenith]

yo how can I get gpt4o on android

[maiden fractal]

You wait till available

[tight zenith]

it is not availble on computer as well

[maiden fractal]

Indeed

[tight zenith]

like I went to the website and gpt 4o was not there

[opaque oxide]

still, its a bug calling the function interface.

[maiden fractal]

It's 4k token of context. Try smaller chunk of code

[inland heron]

The only thing I found strange is that conversations start to deteriorate pretty quickly in complicated code unit-testing. But I noticed this with GPT4 as well, and it happened as soon as they introduced Memory and expanded the output window (or at least that's what it looked to me). It doesn't typically make mistakes, but it can become redundant fast and just goes in circles. This should be fixed, and I just starta new conversation once it happens. Context window is huge for input, so it's generally not a big deal

[opaque oxide]

A different prompt was causing this issue, i used the gpt-4 sys prompt as it made this issue consistent

[maiden fractal]

Make a shorter pre prompt and make it sure he understand the type of code you want to work with

[maiden fractal]

Maybe gpt 4 when it will be available for you will be a viable option

They started rolling out

[opaque oxide]

i was using open interpreter’s sys prompt. I dont want to fix this issue, already found loads of workarounds, just wanted to share it. Eventually it keeps calling functions though. Just annoying.

[opaque oxide]

Nice

How do they prioritize users?

is it based on # of chats or how long the user has been registered?

[light crane]

Random, but starting with plus users

[maiden fractal]

Random

[opaque oxide]

oh

[silk blade]

if you want to code with free 3.5 you need to feed it the latest version language spec from the language's website. GPT4/o you can just link to it.
every agent i've ever asked for help with code has used a mash of deprecated functions and syntaxes if not fed latest specs and told to use modern ECMA methods, but i've not tried to code with gpt for a while

[opaque oxide]

I used gpt 4 all tools for like 2 days, it randomly appeared in my account and i was a free user. Pretty weird.

[silk blade]

i got 4o today, i'm a free user in the UK

[opaque oxide]

very nice

is the voice feature there yet?

[maiden fractal]

Congrats :3

[silk blade]

i dunno, i just chatted with it 🤣

[light crane]

You mean you used the audio and video?

[opaque oxide]

no no

[maiden fractal]

I have access but I didn't play much with it yet.

[opaque oxide]

talking bout old gpt4 all tpols

[light crane]

Oh, okay

[opaque oxide]

like months ago tech

would be crazy if i got access to that tech

[silk blade]

no just text, i'm on pc, i think the speech thing is only for smartphones?
it said it couldn't infer any new scientific conjecture because it doesn't understand causation, so we set about describing a secondary cognition layer that might help it, and it wrote code for it, possibly giving away a bit about it's own code in the process 🤣 predictive text yearns for awareness

[zenith raven]

Took 3 separate prompts with 3-4 regenerate prompts for me

[silk blade]

i wanted to know how it reconciles multiple input types and it seems to be a superpowered version of the 'convert to path' tool from photoshop, converting inputs to a common vector format it understands in tiles. clever stuff

[zenith raven]

GPT-4o is just a cheaper and limited version of GPT-4

I'm debating of subbing for a month to try it and figure out the limits with sub

[glass fulcrum]

80/3hours is GPT-4o limit for plus subscribers

[restive wagon]

Can we get back the "almost OLED dark" mode on the website?

[zenith raven]

Ah OK. Not the worst so

[restive wagon]

plus it also counts with the 40/3h old GPT-4 Turbo

i think this is more than 1 message for every 2 minutes

[zenith raven]

I'm gonna sub for a month and see how it goes. Since I live in ROI, €22.72/mnth (US$24.60 cause +23% VAT) isn't bad-ish

Just to try. I can cancel anytime anyway

[glass fulcrum]

i tryed for 1 month when plus camed out and didn't unsubed to this day

[full sage]

Hey guys , is gpt4o available for free users? I'm confused since they announced but my friend cannot access since he has no pay version

[glass fulcrum]

its still rolling out

your friend just need to wait a bit more

[dense halo]

so many people use it that they have to limit im sure

[full sage]

Got it thanks

[cobalt idol]

limited?

[full sage]

Also any groq user here that can answer a simple question of mine on dm?

[restive wagon]

#ai-discussions

[dapper isle]

can someone tell me why in the browse after some few usage it lags so much?

[daring grotto]

its laggy when scrolling too

[queen kite]

Hey! I just reached the message cap, but it doesnt tell me how much it is? I’m on plus membership, is it still 25 messages per 3 hours (even with 4o model ?)

[glass fulcrum]

Gpt-4o got limit of 80/3hours and GPT-4 got limit of 40/3hours
that limit is dynamic so it can be lower when there is high demand

[queen kite]

Thanks a lot for responding!
But Ugh the limits are getting on my nerves honestly… i rarely use chatgpt but when i need it, it’s for voice conversation and the cap is reached so fast it’s really annoying! They should make it a monthly limit honestly…

[feral zinc]

its hourly because it distributes the server load more evenly. monthly would mean the servers crawl on weekdays/weekends, prime times, etc.

[queen kite]

Yeah i understand.. but my voice convos are for learning languages so we roleplay as if i was a customer or else, so it’s very short answers and messages and the cap is reached so quickly it’s so dumb… while other ppl generate thousands of tokens

Or they could do daily limits idk

Or tokens limits…

[maiden fractal]

Some people switch to api. The limit is your wallet

[feral zinc]

with the current limit its possible to price the subscription to be so low as people cant use all the messages

[queen kite]

API is not super easy access like chatgpt really…

[queen kite]

I’ve tried and i’ll probably try again.. because paying 20$ a month for a little voice conversation and reach the cap so fast is NOT helping lmao

[little cedar]

What is the limit for GPT 4o for free users?

[feral zinc]

currently something like 5-15 messages per day

[little cedar]

day?

[maiden fractal]

Fair enough imo

[little cedar]

cooldown 3 hours

[feral zinc]

yeah ive seen people say they get 24h or timer or something when hitting the limit

3h is for plus/teams users

[little cedar]

no lol...

[feral zinc]

well thats what people tweeted or posted here and reddit so far

[young basin]

Please, try to ignore anything that doesn’t link to a source. Since the announcement I’ve read so many straight up incorrect claims. All the info is on OpenAI’s website

[feral zinc]

and the official link mentions 3h only for plus/teams and no specific time for free users https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4-gpt-4-turbo-and-gpt-4o#h_f6a24d9dd8

[cinder bison]

any idea when the desktop app will be released?

[feral zinc]

"We're rolling out the macOS app to Plus users starting today, and we will make it more broadly available in the coming weeks. We also plan to launch a Windows version later this year."

[dense halo]

later this year part scare me haha. not a few weeks or this summer but year 😭 over reaction i know 😂

[young basin]

https://help.openai.com/en/articles/9275245-using-chatgpt-s-free-tier-faq#

You can see it says 3 hours here

[knotty willow]

Where did you get that from

[lyric token]

I still don't have gpt4o

[glass fulcrum]

its rolling out to every free user

just be patient and wait a bit

[feral zinc]

its weird since i dont believe tens of people faked screenshots and messages and the other page does not list the value

[jagged pawn]

Should it be accessible if you have / buy plus

[feral zinc]

from the release blog

[glass fulcrum]

yes but for free users its still in roll out

[jagged pawn]

Ok 👍

[knotty willow]

Okay mb does it have the new voice feature yet ?

Or just like the web version inside an app?

[feral zinc]

"We plan to launch a new Voice Mode with these new capabilities in an alpha in the coming weeks, with early access for Plus users as we roll out more broadly."

[knotty willow]

Alr thanks

[young basin]

Strange. Got a link to one of the screenshots?

[lyric token]

Well I don't have it

[glass fulcrum]

yea wait a bit

try logining in diffrent browser for example

well or on diffrent device

orr update android/ios app

[young basin]

I can only think that because of overload they are changing the limits for free users until things settle. But when they do, it’ll be 3 hour wait

[lyric token]

@glass fulcrum I'm in eu

[glass fulcrum]

im in Europan Union and for me my free account got it and my plus account too

my free account got it tday

[little cedar]

gpt 4o is poop

[cinder bison]

great, waiting patiently then, i downloaded the DMG from the community but when i login it says i do not have access yet being a plus member

[glass fulcrum]

hm reason ?

[little cedar]

cd

[lyric token]

I did and nothing happened

[little cedar]

24h

[glass fulcrum]

hm try login in webrowser

[glass fulcrum]

its just high demand

you can allways give more H200 if you want oai won't be angry at you

[lyric token]

Done

[glass fulcrum]

works ?

[lyric token]

No

[glass fulcrum]

oh well... if you need something now i can run gpt-4o for you but yea just wait its still in roll out

[lyric token]

I don't need anything thanks anyway

[glass fulcrum]

ok np

[serene horizon]

How will we know in the Android app if we have 4o? Mine just updated but not sure anything seems different

[rare wave]

I think the update is more UI related, the model is more account based

you will get it in all devices instantly

[serene horizon]

But they said in the announcement they will be rolling it out over the next few weeks

That just confused me because it sounded like it would be happening over that period as a whole.

[solid bronze]

I assume they will drop an announcement when plus users have access to voice and vision with 4o.

[solid bronze]

You don’t have to drop down box at the top off your screen in the app ?

[acoustic forum]

I looked through their "capabilities" shown as examples and can't get anything vaguely similar. Even copy and pasting some prompts, things like text generation are not clear and precise as shown.

Are all of these examples custom GPTs built on this model with heavy prompt engineering before it?
https://openai.com/index/hello-gpt-4o/#:~:text=Explorations of capabilities

I can't get remotely similar caricatures. I can't get text to be accurate, etc.

[opaque oxide]

wait, image gen rolled out?

[serene horizon]

I'm so used to not hitting that that I forgot I had it. Just tapped it and turns out I have 4o!

Thanks for reminding me.

[opaque oxide]

”With GPT-4o, we trained a single new model end-to-end across text, vision, and audio, meaning that all inputs and outputs are processed by the same neural network. Because GPT-4o is our first model combining all of these modalities, we are still just scratching the surface of exploring what the model can do and its limitations.” End-to-End, meaning it can generate text,audio and vision?(only images or videos? idk, probably just images if this is true)

[solid bronze]

Nice

[eternal rock]

`` Server Tip: ``

Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!

[acoustic forum]

I reread the disclaimer at the end and just saw "starting to roll out". So maybe it's not ready yet.

[turbid field]

"You've reached the limit of the GPT-4. New responses will use GPT-3.5 until your limit is reset after 22:49."

WHY?

[cosmic hatch]

is this a new annoying feature where you are in 3.5 and it suddenly switches to GPT-4o then tells you "You’ve reached your GPT-4o limit. Responses will use GPT-3.5 until your limit resets after 10:45 PM." and prevents you from continuing the conversation that you thought you were having with 3.5? I am non subscription atm, and this is annoying when it just slides in and ends the conversation, like yes I am aware that my GPT-4o limit has been reached, thats why I'm in 3.5 until it comes back up in a few hours.... "You've reached your GPT-4o limit. You need GPT-4o to continue this chat because it uses tools." this one too XD

[turbid field]

It just dont see open ai

boring... why open to us then?

[thorn saddle]

Guys, Has anyone tried to replicate the content of the press conference? I tried it, but why did I end up with a result very different?

[empty jungle]

So i get the message that i reached gpt4 limit even tho im using gpt4o. Anyone else have this?

[humble pagoda]

It counts for both I think

[empty jungle]

Weird thing is. I can just continue using 4o after the notification. Seems the limit notifications arent working properly i think.

[hazy gale]

because you dont have the updated video and voice

[thorn saddle]

how to update the video and voice?

[hazy gale]

yoou cant you have to wait a couple weeks for it to roll out

[thorn saddle]

oh, I see, thank you

[summer tide]

Ok so there is limited access to chat gpt 4o what does it mean please? How many messages do I get? etc?

it doesn't really specify

[rare wave]

does anyone have vision and new voice yet?

I can't wait 😭

[candid pagoda]

I'm using GPT4o on my iphone with Pro account. I asked her to analyze my voice and she said she can't hear my voice only read what I say which is transcribed.

[oblique plover]

Anyone else get the new Connect Apps when they opened gpt this morning how does it function??

[thorn saddle]

Yes, I tried GPT4o as well, nothing working as shown

[oblique plover]

I clicked the attachment icon and just pulled up my file explorer

[feral zinc]

the new voice features are not yet available

[atomic turret]

When will they be available?

[young basin]

‘Over the next few weeks’

^ from their announcement video

[undone barn]

Same Here!

[knotty willow]

Can we use GPT4o in a gpt yet?

[light crane]

A select few people will be getting it over the next few weeks

We won't get it until after that

[knotty willow]

so much hype fr

[light crane]

Not yet

[warped rapids]

Guys, is this real? "You’ve reached your GPT-4o limit. Responses will use GPT-3.5 until your limit resets tomorrow after 10:10 AM."
I thought the limit would reset every 3 hours.

[novel linden]

I'd imagine its getting hammered more than they expected

[light crane]

I mean, it is free

[spiral siren]

Hello guys, Ive had many slow performances on gpt-4 , is it from the fact that ive havent delete any chat from Janauary until now ?

[cedar glade]

Really? Are you using the free or paid version?

[hearty crescent]

... umm.. so, did they removed the AUDIO CHAT for 3.5 ?

[feral zinc]

lots of free users have reported that the cooldown is 20-24h

[balmy jay]

what happened to chatgpt ? Didn't use it for a month and it is much smarter

[honest crown]

yeah apparently it knows now some of the latest gma shows, 4o rules!

[solid bronze]

Let’s see if 4o can write Code for ninja trader scripts. Because I kept getting errors with 4.0

[urban cipher]

Custom GPTs will stay exclusive for plus members right?

[solid bronze]

No free members will have access now

[urban cipher]

That’s actually really cool, it will roll out in a few weeks right? Or so some people already have access to the free GPT4o?

[tepid patrol]

Hi guys. Since I do not see it written anymore, which is the limit every 3 hour for Plus users?

[cobalt idol]

I know one guy was able to select the gpt4o inside the conversation below the text
while he didnt have the option top left

[tawny edge]

I saw that the new GPT-4o model has become available, are custom GPTs using this model now, or are they still on the old GPT-4?

[dense halo]

old gpt 4 but will update in a few weeks to 4o

[rough stirrup]

how much better is gpt4o to gpt4?

[grim swan]

Is gpt4o out to users or not yet?

[urban cipher]

It’s out to paid users, free users I’m not sure

[glass fulcrum]

rolling out to free users too

[tawny edge]

I've only done a small test but it's definitely much faster.

It's about as fast as 3.5 is now

[zealous canopy]

Is there any estimate on when the image generation improvements will start rolling out? I want to try things with the consistent character generation, and all that.

[tepid patrol]

?

[jovial dew]

why does gpt give way too much answer if you ask something simple. Even if you ask for one line it gives you 57 lines of explanation

[elfin birch]

Working fine for me https://chat.openai.com/share/c4a19b4f-4fb1-44ea-9231-33403296c250

[jovial dew]

yay for you

[elfin birch]

🥳

[mint fulcrum]

nih

[little grove]

In regards to paying for Plus what does "Up to 5x more messages for GPT‑4o" actually mean? I just got access to GPT-4o and reached my limit after about 5 messages.

nvm "Plus users will be able to send up to 80 messages every 3 hours on GPT-4o"

[frigid cove]

must mean that a free user could have sent only 1..,.
check the channel. You will see that 4.o is far from effective...

[winter pulsar]

Hey gang, any of you lads know of any potential legal issues with having gpt to transcribe recordings

[frigid cove]

from a legal point of view, I would say that you must refer to the laws of your own country. In mine, you can be sued for such behavior IF you don’t have the permission to use what someone said.
One of my professors, when I was studying law, had sued one of his students for this...

[mild valley]

Just got the new chat interface. Not being able to edit previous messages is a huge regression imo. Is there any way to revert back to the old interface?

[proven sluice]

Yo, hey guys! Can i use the new voice chat feature already, or is it still rolling out?

[hot crater]

Are you sure it won't work? If I hold the mouse pointer over my message, a pencil icon appears to the left of the text and I can edit my message

[hot crater]

That will come in the next few weeks

[swift osprey]

The UI updates sure are unimpressive

I just want to know when they will ditch the essay style responses from GPT. So annoying to deal with

[mild valley]

Gotcha, I see that now. It seems to not appear on the initial message of a chat if that message has an attachment.

[granite harness]

Yeah, can't edit messages with attachments, but others are still editable.

[mild valley]

I tested it again, maybe it's only if the initial message as an image attachement.

[twilit jetty]

Is GPT-4o as limited as GPT-4 on a Plus subscription?

40 messages every 3 hours?

[honest jay]

It's 80 for 4o! https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4-gpt-4-turbo-and-gpt-4o

[twilit jetty]

Thank you!

[languid lance]

anyone know how can i check how much time i have left untill my cooldown is up?

[feral zinc]

try hovering over the model name

[languid lance]

thank you

[frigid cove]

i encounter issues every times I try to ask ChatGPT to adjust an image.
I get this message with EVERY tries: "I encountered issues while attempting to generate the updated image with red eyes and a forest background. Unfortunately, I was unable to produce the image at this time. Please feel free to submit a new request or modify the current one if you'd like me to try again."

Am I alone?

[visual bloom]

Too much in one go. Take step by step. 1st step - red eyes. 2nd step- rest.

[echo plinth]

https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4-gpt-4-turbo-and-gpt-4o do we now have independent rate limits for both GPT4 and GPT4o?

so overall 120 messages per 3 hours?

[frigid cove]

I would not be so happy... check what is 4.o actually, and come back. Results with it (but without any custom instructions so it could be a part of the problem) are really bad.

@visual bloom ty... I don’t think it is really the problem... I spammed and it succeeded to generate an image. I had aklkso the error while asking fgor a fresh new image.

[raw tangle]

Does anyone know if there has been any discussion about the upcoming desktop apps having any access to local file systems? Even in a limited sense would be nice. For example a specific folder that a user could place files in that the app could access and edit.

[half iris]

Does anyone have access to the app? I received the link to download it but I can't use it yet

[haughty forum]

It appears to just be vision and clipboard.

Settings access has been talked about for windows copilot but I don't think they want to take the risk just yet.

People have done their own RAG projects for local filesystem access though

But that wouldn't tie into the voice model without API access to the voice model which may not come any time soon.

[raw tangle]

Ok thanks. Im mostly hoping for a LLM capable of executing and debugging its own code. I've tried local models but have trouble tying in LTM agents via api.

[haughty forum]

There's some GitHub projects that replicate Devin put there

[raw tangle]

Ill have a search, thankyou 🙂

[haughty forum]

OpenDevin perhaps

You might be among the first to try such repos with GPT4o and given how much faster it is and how well it scores on benchmarks it may be a good option.

[sharp narwhal]

So the announcement says ChatGPT Plus had x5 limit now. So much more usage limits. But I couldn't find anything that says ChatGPT Team limit was changed any. Team was roughly double (per user) limit. But if Plus now has a x5 limit... and Teams didn't get a limit increase...

I might have to cancel Teams and go back to Plus

[kind depot]

when was this announced?

That's what I was just trying to find out.

Does the Plus membership currently give you 80 uses per 3 hours?

[buoyant rampart]

for ChatGPT-4o, yes

[foggy grove]

yeah somthing like that im pretty sure

[kind depot]

oh its so bad

[foggy grove]

huh wdym

[kind depot]

it was double that for the teams, wasn't it?

so logically 80x2 = 160 per 3 hours

[foggy grove]

i dunno i dont use teams

[kind depot]

Do you think 80 messages is enough?

I've been analysing the Chatgpt membership for a few days, but this is the part I don't like 80 messages in 3 hours is a bit ridiculous.

[haughty forum]

My productivity is gonna be awful while I wait for this to launch.

[honest jay]

Just to clarify: the "x5" bit is in relation to the Free limit. Plus accounts have up to 5 times the usage cap than Free accounts. Team is still described as "The GPT-4 and GPT-4o message caps for a user in a ChatGPT Team workspace is higher than that of ChatGPT Plus." https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4-gpt-4-turbo-and-gpt-4o

[buoyant rampart]

Considering it's 40 per 3 hours for 4, 80 is rather nice

[haughty forum]

I think the media is going to throw it under the bus day one like what happened with bingchat

[kind depot]

this rate is too low

[glass fulcrum]

1 message per 2 minutes and you shodn't hit it

[foggy grove]

yes

obviously more would be great, but 80 is plenty, better than opus

[kind depot]

mate, that's impossible. so personally I borrowed it from my friend and used all 80 messages within 1 hour

chatgpt plus membership is unreasonable for intensive use

[buoyant rampart]

It costs a lot to actually run this kind of product

[kind depot]

I don't think so, bing gives unlimited and it's free

[surreal moss]

80 messages Is more then enough its crazy

[foggy grove]

jesus christ wth are you doing with it 🤣 , i use it for all kinds everyday and am yet to run out now with the 80

[surreal ermine]

chatgpt is down rn

????????????????????????????????????????????????

[foggy grove]

right

[foggy grove]

up for me

[kind depot]

Brother, it really is. I mean, I think it's because I've used too much, maybe.

[buoyant rampart]

Same!

[surreal ermine]

brotha

[buoyant rampart]

I've yet to run out

[hazy sphinx]

If you were to use the API to the level of the ChatGPT Plus limits it would cost much more than the ChatGPT Plus subscription.

[buoyant rampart]

Bing probably has ways to subsidize the cost so that's why it can offer unlimited

[hazy sphinx]

Bing itself is a loss leader, no?

[foggy grove]

microsoft has some DEEP pockets too 😂

[buoyant rampart]

Yeah, and I'm sure that offering unlimited is why it is a loss leader but there's also what Speedy just said

[foggy grove]

They can afford to subsidize services like Bing to drive user engagement and collect data to improve their products.

[buoyant rampart]

exactly!

So 80 is more than fair

[eternal spoke]

What’s the coding like with GPT-4o? A lot less errors?

[hazy sphinx]

Mostly just faster for me so far, subjectively.

[buoyant rampart]

It's also much better than what free users get considering the longer wait when the servers are under heavy load

[surreal moss]

what are you doing thats making u run through 80 generatinos per hour?

im curious

[buoyant rampart]

I'm not! 🤣

[surreal moss]

oh im talking to the wrong person whoops lol

[surreal moss]

what are you doing thats making u run through 80 generatinos per hour?

I need to know cuz thats crazy

[buoyant rampart]

lol, all good but I was wondering the same myself

[foggy grove]

fr 😂

ive been trying to hit the limit but aint yet

[surreal moss]

bro must be asking gpt to write his name over and over

[buoyant rampart]

I can barely hit that limit

[foggy grove]

and i feel like if i were to hit it id only have like 30 mins before it resets, cuz with 40 id have to wait like maybe an hour or so after extensive use before they refill

[surreal moss]

yea I would hit the 40 limit often

[kind depot]

I use it for software development.

[surreal moss]

but 80 has been kind to me so far

[foggy grove]

is it possible to do a teams account with yourself and just switch to the other account when you run out if your running out so quickly ??? ( I have no clue if this works )

[buoyant rampart]

Yes, others have mentioned being able to do that

[full veldt]

Honestly the 80 limit seems reasonable - though if you use it as intended (ie conversationally) it would be somewhat low

[foggy grove]

yeah tbf im curious how itll be when we actually talk to it, will the limits work different orrrr ??? So many questions

[radiant kindle]

Is gpt4 free now

[buoyant rampart]

I've been wondering that too though I'm less likely to use the voice feature, severe hearing loss, but it might be helpful for pronunciation from time to time

[radiant kindle]

Can ChatGPT identify music if you play it to it

[foggy grove]

damn that sucks man, understandable tho, yeah I cant wait to use it ive got so many questions and thing i want to test it with

[foggy grove]

with the new voice yet it most likely can

[kind depot]

yeah but only 10 message free

[foggy grove]

I mean it can literally sing to you too 😂

[radiant kindle]

Why so slow?

[kind depot]

welcome to openai free services

[buoyant rampart]

I'm often asking my Dad how to pronounce some words in Japanese since he's spent more time there

[kind depot]

[buoyant rampart]

Because the infrastructure costs too much that if they let billions all over the world use it freely as in unlimited or even higher, OpenAI would have to call bankruptcy and shut the whole thing down

[radiant kindle]

Make it 20 messages than 10 I think that would be more beneficial for everyone and maybe handle by open ai

[buoyant rampart]

They probably already did the math to figure out what they can reasonably allow

[buoyant rampart]

The fact that it'll be better with foreign languages excites me

[blazing hill]

It seems chatgpt-4o is not working in Playground, keeps saying im a ai text based model

[frigid cove]

If I could decide, there will be 0 message free and no free users except by tierce company (bing by example) or as a free trial period

[kind lava]

Bruh

Is gpt not for iOS 15

[proven umbra]

Have they removed the voice feature from the free subscription on Android?

[frigid cove]

seems that there is no more voice feature. You just have to wait the entire deployment

[foggy grove]

i still have mine, im on ios tho dunno if its different

[buoyant rampart]

I don't think the new voice thing has been rolled out yet so it's still the old one

[foggy grove]

yup

cant wait for the new one

[buoyant rampart]

I'll be curious to see how my dogs react to it 🤣

It'll probably scare my girl Schnoodle for sure though

[foggy grove]

Schnoodle 😂

[proven umbra]

I don't have the headphones icon anymore

[foggy grove]

hopefully that means rollout wont take toooo long for new voice

[buoyant rampart]

Oh yeahhh, they're a great breed and it is a fun word to say!

[empty jungle]

Just got an update on android. No update notes tho

[blazing hill]

The one thing I hate about this model is that the answer is always broken. I think it's intentional so that it doesn't put much effort and energy in answering

[buoyant rampart]

Yeahhh, I had an update yesterday for iOS but nothing notable that I could see so think they're just doing bug fixing for now

[blazing hill]

I always have to type continue and it's frustrating

[buoyant rampart]

Huh, that's a shame to hear. I haven't run into any issues with it so far other than it hallucinating about a couple things

[blazing hill]

It's only useful for short questions or prompt requiring short answers. It favors accuracy but at the cost of length

For example I asked for a movie synopsis in 10 chapters it gives 2 and 3 is broken

[dense halo]

will 4o gpt be in copilot

[little dock]

how to use gpt40?

[swift osprey]

When will the essay style responses end.... It's the worst thing about GPT right now

[swift osprey]

should be a dropdown menu at the top of the chat window

[little dock]

is not free?

[swift osprey]

Ah I have heard it is free. Idk how to access for free users. Maybe its the new default?

[little dock]

ok

ty

[cold pewter]

In training chatgpt to get to know me better through memories, i cant find any of the memories from yesterday, just rhe new ones from today

[boreal lily]

i hate how all of my msgs get defaulted to 4o and its eating up limits

[gleaming idol]

How many limits?

[foggy grove]

im glad it does

[cold pewter]

oh... my god... the memory is self deleting... you could tell it all about yoursef for 6 hours and the next day its gone

[eternal rock]

This is a Safe for Work Discord Server

The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.

Posting NSFW content will result in a permanent ban.

[foggy grove]

?? mine are fine ?

[cold pewter]

yah maybe you havent done as extensive of memory testing, i gave it hundreds of memories over the course of 12 hours, it only has the ones from the past 2 hours

it could be overwriting them

[boreal lily]

free limits

[foggy grove]

mine has around 150ish

give or take 10

[boreal lily]

there should be an option where you can set the convo model

[foggy grove]

yeah think theres an option to like default to the last used model right?

[boreal lily]

i cant find it

[cold pewter]

odd, it has completely forgotten all of yesterdays memories. i manually delete all my conversations but that did not matter yesterday

thats a pretty rough bug for an ai assistant lol

seems its a known bug. that sucks, guess i can quit training it for now

[quick jasper]

Hi folks. I must apologize first before I start. This is my first trip into Chat-Gpt4.o
It might be too soon after the rollout but.. I paid for the subscription, and now trying to figure out what Im doing lol.
On the computer, and on my Android Pixel7pro I can see the GPT4.0 but. When I tap the headphones in the btm right? She tells me she is on V3.5 lol
According to her? 4.o has not come out yet lmao

When I text chat with 4.o? She tells me to use text to speak because she can not "voice chat"?
Pease help an old pc engineer (that's new to chat gpt) out of a jam lol

[buoyant rampart]

It's true, 4o hasn't yet rolled out its voice feature

[hard fossil]

The open ai discord app bot just went offline 💀

[native spear]

this bug keeps happening. Is there any way to report this?

[hard fossil]

It's back

[hazy gale]

what are you using the discord app bot for?

[quick jasper]

Thank you! 🙂

[hard fossil]

Nothing Its just one of my most recent chats and it went offline

[buoyant rampart]

No problem! It should be within a few weeks at least

[keen meadow]

This glacially slow rollout of desktop app access to Plus users is pretty frustrating. Anyone recently granted access?

[visual nova]

the app was the part i didnt get that well during the demo

is it just the same thing as the web version in an app wrapper?

Or doe it already have hte voice video in it

[wanton lichen]

I agree, I've been dying to try it for programming but no dice

[keen meadow]

No, it has access to your desktop, loosely speaking. You can quickly get ChatGPT to interact with images or text on your screen, once all the new features have rolled out to GPT-4o.

[visual nova]

oh, like it screen records in real time? I saw that would be possible and wanted to make it in the api.

That's cool, it might mean the compatibility is also harder to get than I thought since I'm on linux.

[keen meadow]

I don't think it's recording the screen constantly, that would be insane. More like Spotlight, and you can quickly interact with things on your screen.

[fickle dove]

Question since I'm brand new, can you use Chat GPT 4o right now with the plus version? I've been seeing it advertised all over Tiktok I was wondering if its out right now to use.

[glass fulcrum]

Yes GPT-4o is live and you can use it with plus or free account (you may not got it rn on free account as its still in rollout)

[fickle dove]

That includes all features advertised? Like voice convos and cameras?

[cedar glade]

Just text like the old models.

They are rolling out the new features they demoed up ahead.

[fickle dove]

Ah I see, well then I guess it’s not worth it right now. Is there a date on when all the features will be released?

[cedar glade]

(or, it's faster and stuff, but it's still the same)

They said "the upcoming weeks", so could be next week or after summer I guess...

[fickle dove]

Thanks for the info friend, much appreciated! 😁

[cedar glade]

👊

[eternal spoke]

Is it me or is 4o slow rn

I noticed in long convos it just freezes and you need to refresh after like 30 secs

[acoustic talon]

Are the custom gpts in 4o or nah?

[modern horizon]

Long conversations don't work well, especially in the mac app. It's real glitchy. Also using 4o.

[autumn idol]

hi guys a question, but gpt 4o being a multimodal model, does it also have the ability to transcribe a proper audio file? i tried to upload the file and ask it to do transcription but it can't return the result.

[native spear]

cant as of right now

gemini 1.5 pro can i think

[autumn idol]

ok, thank you very much i appreciate it

[twilit temple]

What was the eta for the new voice features again?

[feral zinc]

"We plan to launch a new Voice Mode with these new capabilities in an alpha in the coming weeks, with early access for Plus users as we roll out more broadly."

[twilit temple]

Cool beans, thank you,.

[halcyon lava]

Is the new voice mode available in German language too?

[dim aurora]

I’m assuming so, based on the translation in the announcement?

[halcyon lava]

That would be great

[woven juniper]

Hi there, anyone else noticed GPT4o hallucinating with code? I tried attaching a python file (not too long of a context) to find an issue and it came back with spotting functions that were not there in the original. Similar with pasting code in the window. Never seen this before with GPT4.

[jade crater]

Is iOS ChatGPT voice chat freemium like 4o features? Is it part of 4o?

[acoustic talon]

are the custom gpts on 4 or 4o??

[spark storm]

Recently, I saw a video on the OpenAI YouTube channel titled "Live demo of GPT-4o's vision capabilities," where you demonstrated a feature on an iPhone that allows for live conversations and lets GPT-4o see the world around me.

I would like to ask why I am unable to use this feature even though I have access to GPT-4o. Do I need to wait for this feature to become available to me?

[dim aurora]

It’s not fully available yet! It’s being rolled out to plus members (I think?) and will be available in the next few weeks.

Are you a free or plus user?

[spark storm]

PLUS. I have been a Plus user from the beginning and am committed to OpenAI. I like the service and clearly understand the roadmap while using it. However, I have recently started to dig deeper into OpenAI. I am not interested in speculation and misinformation, but rather in specific and constructive information!

[dim aurora]

Sorry man, just trying to help.

[nocturne pewter]

I’m using gpt-4o on my iPhone however when I ask gpt the difference between 4o and 4 it argues that there is no gpt 4o. Any idea why?

[spark storm]

I am grateful and thank you!

[dim aurora]

No worries.

[topaz swift]

is the GPT-4o limit for free users a total limit, or does it reset daily?

[dim aurora]

It resets every few hours I read.

[topaz swift]

Oh phew thanjs

[spark storm]

Since it is still under development and continuously expanding, it is stated in the bulletin that it is smart and fast. Currently, GPT-4 is only smart! So, while the speed might not be highlighted now, perhaps in another iteration it will be.

[dim aurora]

I’m guessing you reached the limit?

[topaz swift]

No I wanted to make sure before trying it at all

[dim aurora]

Oh fair enough

[nocturne pewter]

Thanks

[dim aurora]

I wanted to see if anyone reached it in the shortest amount of time😂

[lethal spire]

The vision should be released asap

it'll be a huge deal for coding

[haughty forum]

Vision is already available, it just won't be available in tandem with the new voice model until.... Shrug

[median meteor]

why i dont get the macos app yet ._.

[elfin birch]

It’s still being rolled out to Plus users

[lethal spire]

How? Can you talk with it like in the videos, but without that humanlike voice?

[honest ledge]

Hey, I have a couple of questions regarding GPT 4o

1. as a GPT plus user, what is my new rate limit, they said 5x, does that mean 200 messages every 3 hours?
2. 4o is supposedly multimodal, but I can't seem to upload images, why? Is this going to be rolled out gradually?
3. will 4o enable free users to access custom GPTs? I built a custom GPT, but haven't been able to have anyone I know test it as they don't have GPT plus

[maiden fractal]

gpt4o is 80 per 3h and gpt4 40 per 3h

[maiden fractal]

gpto still being roll out

and free user dont have access to gpts

AFAIK

[honest ledge]

Ok. How big is the context window with 4o?

[grim pumice]

Hmmm, I got a banner on the site to download the app and I did so but when I try to login it says I can't use it yet. Why did it prompt me and tell me I could?!?

[jolly yew]

guys, what is the limit to gpt 4o for free users? also, does it just switch back without you knowing to 3.5, when you reached the limit? 🤔

[maiden fractal]

few message per hours,

[maiden fractal]

good question ill try to find the answer

[jolly yew]

ah i see.. yo ucan switch model after getting the response. thats a bit weird

[maiden fractal]

can you share the screenshot of the email

[jolly yew]

ouch thats rough. makes me wonder, why not just use gemini

[maiden fractal]

thats #ai-discussions debate.

[jolly yew]

👍

[grim pumice]

It wasn't an email, it was a banner message right above the chat input box.

[maiden fractal]

they are rolling out the app not yet available to everyone

for apple user *

[grim pumice]

Yeah, it's just weird that it offered it to me - 'here, download this now' but then it doesn't work.

[maiden fractal]

are you on windows?

[grim pumice]

No, macOS. App runs fine but when I login it says it's not enabled for me yet. So, some sort of mismatch between who they are prompting to download the app and who actually has access.

[maiden fractal]

as far as i read 32k for plus user

[brazen mortar]

Is the new voice feature out on the app? I have 4o but nothing new about it

[blazing hill]

Chatgpt is so good for trolling

[hushed grotto]

From what I’ve read it’s going to be rolling out over the next few weeks, no exact dates given

[tepid patrol]

Do you guys know which is the current limit for GPT users? I remember 40 messages every 3 hours. What about now with omni?

[honest jay]

4o has a base limit of 80/3hrs for Plus users! https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4-gpt-4-turbo-and-gpt-4o

As of May 13th 2024, Plus users will be able to send up to 80 messages every 3 hours on GPT-4o and up to 40 messages every 3 hours on GPT-4. We may reduce the limit during peak hours to keep GPT-4 and GPT-4o accessible to the widest number of people.

[tepid patrol]

Thanks! I missed that part

[bronze magnet]

Did they remove the chatGPT plus free trials? I remember they gave us 3 links to give friends free trials. I used 2 and had one more but can’t find it in the UI anymore

[native spear]

Just to be clear:

• if you reach 40 messages on gpt4o, you reach the limit for gpt 4 turbo

So basically you can send up to 80 messages in total for all gpt 4 versions. It's not 80+40

[queen rune]

new release 4o bugs:

1. input text box becomes hidden when attempting to edit
2. Edit button moved to top of input box (means the user has to scroll to the beginning of a prompt to edit only to have to scroll back down to the end. Ideally this should be placed at the end)
3. text generation freezes mid-response. Unless refreshed it will remain hung up like this indefinitely.

[foggy anchor]

gpt 4o 🗣️🗣️

[sick pecan]

i got offered a download link for the osx app. only to be rejected at login for it not available to me yet. open(f)ai(l)

[native spear]

Yea not to mention the erasing all memories at once bug

[vagrant ginkgo]

Ya same

[tropic star]

How can I open the camera on my phone just like in the demo?

[sick pecan]

the demo features are hype not available yet

[tropic star]

Ohh, thanks

[sick pecan]

they only did it to mock google fake demo before google got a chancne to redeem themselves

[pulsar anvil]

Is the app out for all plus members or is it batched rollout?

It just dropped for me like just now

[sick pecan]

i think we all got the link notification, but not all of us can actually login

[pulsar anvil]

Yeah that's what I meant

A download popup just showed up on my browser chat and I'm guessing from the openAI post "We're rolling out the macOS app to Plus users starting today" that it's out for everyone?

[solid bronze]

How do you acces the vision on the plus mobile app ?

[haughty forum]

The camera and photo icon in the lower left.

The current voice modality isn't as good but you can enter and leave call mode and it retains context.

You can also ask for images to be generated in call mode then check it later when you leave the call if I remember.

You're sending it the photo rather than saying "what's this" in voice conversation. But it's still vision modality.

[sick pecan]

i got chatgpt for android 1.2024.122 and it still is same old stuff

i also got a play store update pending for it derp

.131 and still bleah

[zealous canopy]

Anyone here got GPT 4o's multimodal capabilities yet?

[boreal lily]

i got gpt4-o and honest opinion
id just stick to 3.5

[zealous canopy]

Hard disagree. But im asking if anyone has gotten its stuff with the voice mode and the image generation improvements yet.

[rustic kelp]

ChatGPT desktop app says: "You do not have access to the desktop app yet. You can still use ChatGPT at chatgpt.com"

pleeeaase, team, don't do this thing Google does where they punish enterprise/business level customers.

just don't release the app yet if it's not fully functional for all users

[zealous canopy]

I dont think they have released it? From anywhere I can find it its obviously some kind of unofficial release.

yeahhhhh @rustic kelp I would run a virus scanner bud.

[urban cipher]

I thought free users will get access to GPTs as well? Didn’t they say this during the live stream? But idk

[shell topaz]

Does anyone know if the upcoming update for the voice feature will improve how it sounds in other languages like dutch? In the current version dutch voice sounds quite broken

[urban cipher]

Are there people already having free access to GPT4o?

[raw lance]

So free users can have unlimited gpt4 on phone, is it true? But on desktop it's just limited gpt4o

[maiden fractal]

some yes

[maiden fractal]

free user do not have access on gpt4 only gpto and limited

[raw lance]

But I ask GPT, it says it's gpt4 and unlimited. I guess it's confused too

[urban cipher]

Lmao I just got free access as well

[raw lance]

it's not called free access, but it's trial access. they want you to pay for plus.

[empty ice]

how to get access to gpt 4o

[urban cipher]

Yes you have access to the new GPT 4o model, not the old GPT 4

[empty ice]

how

[urban cipher]

Oh fr? But the Limit resets right? Or is it a one time thing

[empty ice]

i only have gpt 3.5

[raw lance]

reset at every 6 hours or more? and not sure if it's the same times of goes after reset

I have tried it's document reading capability, and it's aweful

[urban cipher]

They have, I just tested

[urban cipher]

Yh as always in GPT4

[empty ice]

this poe website from quora has access to gpt 4o

[raw lance]

True, I assume because everybody has aweful feeling towards their Plus service, that's why they want to rename GPT4 to GPT4O, but honestly nothing changes.

[urban cipher]

It’s just faster and a bit better structured outputs, but yes most is the same. And the trash document reading is really bad tbh, it would be so much better if this improves. Sometimes it just denies that some information is in a document

[raw lance]

I guess many people have used Bing Chat, based on GPT4, its document reading and image creation are much better than ChatGPT Plus

It's like they want to provide aweful service on purpose. As many people have discussed here, quantity doesn't secure quality. In this case, less quality.

[raw lance]

Absolutely

[odd flint]

Chatgpt Voice is not working right now. Can someone confirm it to me ? Everything was fine 4 hours ago. Are they upgrading, or messing with it ? 😮😮😮

[urban cipher]

An Hour ago it was fine with me

[odd flint]

Looks like it's offline. On Android phone app, at least, I can't talk with it using voice. It tries to connect, but fails... 😌😌👍

[devout jetty]

Anyone here?

[urban cipher]

Idk I’m on iPhone, has it disconnect once but it reconnected and works now

[empty ice]

how do i get access to gpt 4o

[urban cipher]

Either you buy plus or you wait until it’s available in your account

[odd flint]

Interesting is that OpenAI is launching talks with voice as something new. I talk to Chatgpt Voice for at least 1 year now. Perhaps some people didn't even know about it. I almost never type anything anymore. We talk hands free. Now, they are presenting the new Chatgpt 4o as something completely new. It's just an update for the voice, with more inflections, understanding, but voice already existed, and it is awesome 👍. 💥💥💥

[empty ice]

some free users haave access tho

[devout jetty]

That is why your typing speed is slow

[sick pecan]

fast voice with interruption is a completely new experience, if only incremental technical updates

[odd flint]

It disconnected for my account too, few minutes ago. After almost 1 year I had to login on OpenAI account again. But exactly after that, the voice doesn't work anymore. And there's a new message from OpenAI saying "Now, Chatgpt has voice"... Oh my... It already had it. I use it for more than a year. 👍

[sick pecan]

is a major milestone of the natural language computer interface revolution

[odd flint]

My type speed is slow perhaps because I use only one finger to do it... Besides, I don't know English. It's my third language 🤣😅😅

[stoic coral]

any way to get the mac app any faster?

[elfin birch]

Speed up time

[odd flint]

I'm a plus member. On the phone app, there's a new Chatgpt 4o, but for now, nothing changed from the 4 version. At least for me. 👍

[devout jetty]

This content is blocked by this server. From server moderators:
‘Non-whitelisted URL or slash command detected. Your message was not sent.’

[urban cipher]

Yes that’s why I’m saying wait until you get access, I got access an hour ago, i think it’s regional

[devout jetty]

Huh

[urban cipher]

Limit for free users is 16 messages btw

[devout jetty]

Why is DALL 3 not working

Instead I got a timeout

[glass abyss]

Anyone else noticing that the training data cut off is different each time you ask. Yesterday when the model released it knew about the latest version of Angular (Code Framework) Today it has no idea.

[odd flint]

Have you access now to 4o on your phone ? Do you use Android ir iPhone ? Perhaps it's regional. I live in Brazil, I hope they release for me also ( plus ). 👍🙏

[odd flint]

BTW, the worst thing for me about Chatgpt 4 or any other is the message cap. In the best of a conversation, using voice, then comes a new voice saying: "You've reached the message cap..." Oh my ... 😮

[mild cave]

yes. its all over the place

[glass abyss]

what a mess, they always do this. Release a really good model and then destroy it.

[toxic lake]

is gpt4 any better than 4o?

[glass abyss]

no

at least for what I do

[elfin birch]

Try them out and see how they work for your use-case

[toxic lake]

i use gpt for everything

nothing specific

[elfin birch]

GPTs currently do not use GPT-4o

[dense halo]

you guys see this? wow. i cannot wait 🤯 https://twitter.com/gdb/status/1790869434174746805?s=46

[toxic lake]

im asking if 4 is better than 4o

or is 4o just superior in every way

[elfin birch]

I personally think GPT-4o is better, but it really depends on what you're asking it to do

That's why I asked you to try it

[dense halo]

you have to wait for the next update for people to say gpt-4o is best. internet always says 'i think the last model was best'. gpt4 was better than turbo to some, now turbo better than 4o. it is called contrarianism

[rocky wing]

I can certainly tell that 4o is faster than 4. Is it any smarter? As in, have they improved the quality of responses?

[toxic lake]

3.5 was never better than 4 objectively, not sure how you can compare them

[frigid cove]

for now, the interest of 4.o is that you can use it when you have burnt all tries with GPT 4

[prime swallow]

That's really going to depend on what sort of question you're asking it. To my testing, the stuff 3.5 was really good at, it's still really good at and 4 and 4o isn't that much better. There's a range of stuff that 4 can do that 3.5 really can't, but I haven't found something yet that 4o can do well that 4 cannot do well. The outputs differ, but I like them both and happily use them both.

[sick pecan]

3.5 can be better if you have a use case that works with fine tuning and requires cost efficiency

[dense halo]

Is talk/voice mode not working for anyone else? It says connecting and spins for a little bit before booting me out. Sorry if it's been asked already.

[round sun]

Did it update for you?

[dense halo]

Personally ive been using exclusively 4o and I've found that the responses seem to be noticeably higher quality than 4

Could be in my head though

[round sun]

Any news for the free tier regarding 4o?

[dense halo]

Yes, it's fully updated. I just downloaded the app on my phone to try it out since it was mentioned on the release of 4o. I have premium, so I should have access to the feature as well.

[round sun]

Oh, ok

I can't even see the headset icon

I could before

Maybe they disabled it because too many people are using that feature?

[dense halo]

I have access to the voice-to-text option which has been working fine but the headset one just doesn't work when I click it. I think they might be rate-limiting/disabling it because of high usage with the release of 4o

[round sun]

Or maybe because they're slowly rolling it out

[broken field]

Is the Mac app available yet?

[light crane]

I believe so

[broken field]

I don't have access yet :(. Probably no way for me to get access until they message me or smth

[light crane]

I've seen people download it from somewhere, idk if OpenAI themselves uploaded it

[broken field]

oh ok, thanks

[gaunt gyro]

my chatgpt is very slow for the past 4 days. anyone experiecing it? how to fix it?

[fervent agate]

It will be. New model release results in significantly increased traffic/server load. Not much you can do until it dies down.

[upper arch]

In ChatGPT free version, when I'm creating a new chat, the system will automatically use Gpt-4o as the default. Is there no way to make it force-select 3.5?

[burnt hare]

no, but you can bookmark the new 3.5 chat link, the model is in the url now i think

Does anybody with a Teams account have a few moments

Are you able to eg create multiple "project"s or "workspaces" or what have you and then create chats under that unit? does it provide better organization?

[glossy fjord]

why didn't GPT restart after 2 hours and it says that I have to wait 24 hours to use GPT 4o again?

[swift pond]

if you are a free user, it might be caused of high usage in general by everyone, but that's just a guess

[smoky pilot]

Been beating my face against the wall all day trying to use Chatgpt to help with Python and openai. It hasnt gone well

[lunar lichen]

hey guys i have a question. I believe chatGPT desktop app for macOS is not available to Team plan users am I right?

[swift pond]

they are interested in feedback about the performance of the model https://openai.com/form/model-behavior-feedback/

We’d love to hear examples where our models did not meet your expectations. Sharing submissions helps our research team improve our models.

[autumn perch]

voice is not working here

[rancid otter]

what does this mean? I just normally chat with it

Our systems have detected unusual activity from your device. Please refresh your browser and try again later.

[rustic kelp]

nope. https://help.openai.com/en/articles/9275200-using-the-chatgpt-macos-app

[zealous canopy]

I dont see an official download here. I hear its supposed to be downloadable from the webapp.

[rustic kelp]

it was. it popped up & told me there was an app and it let me download it

[zealous canopy]

oh in that case youre good.

[empty ice]

GPT-4o is available in poe

from quora

[wise yacht]

Just got access to the Mac app. Didn’t think I’d care as the website works fine, but this is quite useful.

[zealous canopy]

do they even have plans for a windows version?

[rustic kelp]

so sad it doesn't work for me 😦 it let me download the app but I can't use it yet for some reason.

[wise yacht]

Are you using Sonoma? I had to update

[rustic kelp]

I'm updated. it says "You do not have access to the desktop app yet. You can still use ChatGPT at chatgpt.com"

[wise yacht]

They do indeed. When, I have no clue. But they did say a windows version was coming

[weary charm]

Can you use gpt 4o on phone browser

[wise yacht]

Ugh. Frustrating, sorry. OpenAI rollouts of new features is bizarre

I waited months for Memory function lol

[thin isle]

Yes!

[honest tide]

where did you get the app Sam?

[woeful vigil]

is anyone else unable to connect to voice chats with any model on the iOS app? it always shows the buffering icon, then "connection failed"

[swift pond]

I've seen a few people mention it, if you search "voice" you'll see what I mean

[strange ruin]

Has anyone come across whether ChatGPT (especially voice mode) has any sense of time? Seems like a missing piece to human like behavior.

[swift pond]

imagine if it was like "I'm on break" hey wait a minute.... /s

[wise yacht]

Not enough talk about the web search speed of the new model. It’s fantastic. Bye Perplexity it was nice knowing you.

[swift pond]

it is spiffy

[royal viper]

anyone else getting a lot of recent request failed please retry errors on 4o?

[silent ingot]

gpt 4o is soooo laggy after messaging for a bit, anyone else

[eternal rock]

`` Server Tip: ``

Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!

[royal viper]

yup, either that or it errors out eating up my message limits

[silent ingot]

so bad

[jolly temple]

has anyone gotten the video chat thing

[jolly temple]

to be fair, they said theyll release it to "small portion of users"

normally for new features theyll release it to all users within 1-2 weeks

[burnt hare]

whole damn domain down for me now lol

[jolly temple]

fine for me

[burnt hare]

sad trombone

[jolly temple]

[royal viper]

just errors for me

retry on app, unusual activity on web

[burnt hare]

yeah i was in the middle of a conversation and now i don't even get a 500 or a 404

edge just spins

api is up though

[lyric meteor]

Hey guys, sorry fif this has alreadey been asked but is there a confirmed release date for free users obtaining GPT 4o?

[vagrant island]

I don't believe so, I think they're rolling it out

[upper arch]

It rolled out for me as a free user, the only problem is that the model is not user-selectable. Meaning, in a new chat, the system will automatically choose GPT-4o by default. It's annoying because there are cases where I need a new chat, but prefer using 3.5 so as not to use my GPT-4o daily quota.

[burnt hare]

just bookmark the URL of the new 3.5 chat

[upper arch]

yes, but doesn't that remember the previous context?

my purpose of a new chat was to clear the previous context especially when the chat have gotten very long

[burnt hare]

chat.openai.com is not working for me or i would demonstrate but i seem to recall that when you select "3.5" in the menu, the URL in your browser changes to something like chat.openai.com/gpt3.5

once you start a conversation it gets a guid URL

[upper arch]

I actually never encountered that, it had always been dynamic guid URL for me. chat.openai.com/gpt3.5 is 404 .

[burnt hare]

does the url not change when you select 3.5 from an new chat

[upper arch]

https://chatgpt.com/?model=auto

It only shows that URL when I select ChatGPT (great for everyday tasks)

[burnt hare]

well play around with the model param

and know that i am envious because for the first time in months it's down for me

😄

[upper arch]

Unfortunately the system accepts any parameter, even non-existent ones like https://chatgpt.com/?model=3.4 . so testing this will just consume my limited 4o quota

[upper arch]

you said chat.openai.com is not working. What about chatgpt.com?

[burnt hare]

they are both back up for me now

well, they were for a second 🤣

[swift pond]

gpt on a coffee run

[burnt hare]

i think it's AT&T actually

hmph

[hearty ice]

what temperature do you all recommend?

[swift pond]

for what use case?

[hearty ice]

for example, i give it a research paper and i want it to reply to my queries based on that research paper

[glossy fjord]

why didn't GPT restart after 2 hours and it says that I have to wait 24 hours to use GPT 4o again?

why now GPT 4o only 24 H used again ?

[swift pond]

my point is like you might want to use a different temperature in that scenario versus like you're playing around with generating poetry

you can play around with it and get a feel for it

[median meteor]

i dont get macos app yet :c

[median meteor]

yeah, couldt wait xD

[tawny crater]

WHAT IS GOING ON WITH GPT. i am telling it to stop to try to explain something it is doing wrong and it keeps generating text.

I don't mean that

I mean i tell it to stop generating so i can give it corrections. And it ignores it and tries to rewrite it again without me giving it any corrections

It should say understood. Not continue writing

It doesn't matter what i tell it. I literally wrote Hi. And it still did it.

Sigh. Again thats not what i mean.

It is writing the same passage everytime i ask it to stop.

I'll dm you cause i dont feel like sharing with the world.

[topaz elbow]

Finally received the download link for the MacOS app! LFG! But, can't sign in after downloading. Guess I need to wait a while

[dense halo]

Same here. Told me I could download it. Did so. After logging in and proving I’m a human, tells me no, sorry. Maybe because I have a Teams account but notifying me that it’s available and then telling me it’s not just isn’t a fun experience lol

[lusty island]

What is the message limit for GPT-4o for plus users?

Also is the token limit on 4o higher than 4-Turbo?

[dense halo]

In ChatGPT itself, a popup bubble with the download button, so I assumed it was ready but apparently it’s not

[buoyant lodge]

same issue i just restarted my laptop

[west ridge]

still waiting

[inland aspen]

how do i use the multimodal feature for the new gpt-4o

[buoyant rampart]

It's not out yet

Give it a few weeks though

Same! I'm looking forward to seeing some new and unique use cases of ChatGPT

[broken fractal]

sooo... does chatgpt NOT analyze audio, without going into code and failing every time?

[nocturne herald]

for those who tried GPT4 and GPT4o, what's best ?

in term of coding

[mortal steeple]

Its dog slow like slow slow slow

[buoyant lodge]

they are honestly about the same, but I have been using 4o for the speed

[minor niche]

gpt 4o can understand the sign language ? anybody knows

[prime swallow]

Try uploading some images with it and see! We don't have a way to give the model video yet (but that surely comes with the fuller 4o release)

[minor niche]

ok

[light crane]

Actually, that's how long it'll take to stop testing. It will start rolling out after the few weeks

And it's possible by "coming weeks" they meant more than a month

[buoyant rampart]

Makes sense

[dense halo]

wait so gpt 4o is just limited, when is it going to get revoked for free users?

[rocky valve]

Never

[dense halo]

wait

so like free users can use 4o forever

[rocky valve]

Yes

[dense halo]

forever?

[light crane]

That's true

[rocky valve]

Paid just gives you an increased limit

[dense halo]

oh so its just like "youve reach limit per hour"

[light crane]

Per day

[rocky valve]

Yes

[dense halo]

ah ok ty

[light crane]

As of now, for free users

[prime swallow]

Until something changes. It's rolling out, and uses per 3 hours is based on total demand on the system, and will be low. But you can go between the free 3.5 model and 4o freely when you have access to 4o

[rocky valve]

idk what the limit is since I have paid so have never reached it

[light crane]

80

But it varies based on traffic

Sometimes free users get only 6 generations, but it's up to 16

[dense halo]

how do you make it like it use 3.5, cuz its wasting my 4o uses lol

[prime swallow]

Upper left side, there's a way to select which model you use; at the bottom too right above the text box, kinda the right-most option to pick between your model choices

[light crane]

Tap the text that says ChatGPT 4o

[rocky valve]

Create a new chat then click ChatGPT-4o and change it to ChatGPT-3.5

[upper abyss]

There's no way to select 4o now

[dense halo]

yea its just this

ChatGPT Plus(Upgrade)

and ChatGPT

no way to switch

[rocky valve]

Yeah if you have it on GPT-3.5 or GPT-4.o you have to create a new chat to switch

[prime swallow]

You just click what you want to use for your next input, and then type your input and enter, it gets sent to the model you pick.

As a free user, if you have access to 4o they said you default to being on that model - but you can change it before you start a conversation, or during it. And if you run out of 4o uses, it'll automatically just move you to 3.5

[light crane]

Not everyone has it for free

[rocky valve]

Oh yeah it's probably still rolling out for free users

[prime swallow]

Not sure you're looking at the right spot to switch

[upper abyss]

It will work abd the a few hours later my app design Will change and things will stop working

[dense halo]

upper left right?

has temp chat also in it

[bold anchor]

A very tiny update: I’m still on 3.5 but the “Get Plus” icon on the iOS app now says getting Plus will give me access to 3.5, 4 and 4o. Before it just said 3.5 and 4. So now at least my app acknowledges 4o even if it doesn’t have it yet lol

[glossy fjord]

why do I have to wait GPT 4o to 24 hours for GPT when last day the waiting time was estimated to be only 2 hours?

[prime swallow]

Are you a free user?

[glossy fjord]

yes

[prime swallow]

Maybe is heavy use, and not so much available for free users right now. They say that they give access based on how much use

[glossy fjord]

but still yesterday it was a 2 hour wait and today I have 24 hours for the same number of words

[prime swallow]

The difference may be on OpenAI's side, not yours

[glossy fjord]

It's really worth it for us to deactivate GPT 4o because we don't like waiting for something like 24 hours

it doesn't matter, it just motivates me to look for artenative

[buoyant rampart]

Indeed, I think it might be like this for awhile because so many are excited about ChatGPT-4o

[prime swallow]

I'm just a community member like you are. I don't have any control, can't change anything, just I can sometimes help answer some questions if I know the answer.

But! You can share ideas in #1070006151938314300 or even #1070006915414900886 and let them know what it is like, what you like and dislike, and what you think would help you enjoy the service more!

I hope you get the help and access you want 🙂

[glossy fjord]

so they could give the same vote as in the past not this one

(for example, many users use our services or some functions of great interest to users of the free plan cannot be used, etc.)

You have exhausted your GPT-4o limit. Replies will use GPT-3.5 until your limit is reset at 20:22.
11 h done and 13 H wait me reakt 😩

[buoyant rampart]

I'm afraid I can't do much about it, not even as a contractor

[glossy fjord]

yes yes, but it would be wrong if they gave an answer like that, you don't even know how half of the people could lead, fine, it's a burden on the user's interest, but that's the way it is, if it's a deliberate setting like this, and the impression of interest is not equal to the view that has it now

[karmic tartan]

anyone managed to install chatgpt on android tv? really sad that its not supported because lots of tvs now have integrated camera

[glossy fjord]

hey, is it true that apple paid open AI money for the GPT model so that they could put this GPT model in the new apple phone machine or crap?

[karmic tartan]

We will found out this during WWDC next month i guess.

[glossy fjord]

try sharing app your phone on the TV?

I'm just looking forward to it because if it's true, we'll have GPT out on the internet by the end of the year, so there's an 80% chance that someone will break the chip and pull the model out

[karmic tartan]

it would be much better experience with standing in front of tv .

[glossy fjord]

create a GPT package for your mobile, forward the package using the BT / wifi / USB port and install it manually

[glossy fjord]

or even a variant of the browser

[spring creek]

Why can't I access GPT 4 for free? Is it not released to the public yet?

[glossy fjord]

as far as I know, GPT 4o is accessed slowly and not everyone has it right away

[spring creek]

so gradually until everyone can access it?

[buoyant rampart]

That's right

[glossy fjord]

yes, but you have to wait
unfortunately, someone was the first on the waiting list and someone, unfortunately, was the last wait, like me, I have to wait for the new order limit for GPT 4o to 13 hours, so you also have to wait, unfortunately

[random stratus]

Is there no option to select 3.5 model as default? I really can't seem to find any. I already tried the new chat approach.

[prime swallow]

Apparently not as a free user, but feel free to #1070006151938314300 the idea! It's very new that free users could select anything!

[full sage]

does gpt models run 4o?

[dense halo]

in a few weeks they say

now still 4 turbo

[boreal timber]

As longer my chat gets (and it must not be long), inference gets slower and slower until the model (most cases i tried with gpt-4o) completely stops generation... mid inference...
Is this normal for now?

[opaque oxide]

Hey guys, if gpt4o is smaller, cheaper and better(some ways) to gpt4, there is a chance that they pretrained gpt4 for some time, didn’t do any safety or rlhf and made it omnimodal, then distilled that bigger model into a gpt-3.5 size model and then did some safety precautions. This makes the most sense and would mean that there is a bigger and better model behind closed doors

[opaque oxide]

the stoppage is not normal, however it slowing down on getting more input is to be expected when working with tranformers as they scale their time with token’s quadratically iirc

[dense halo]

i would guess there is always going to be a bigger model behind close door haha

[opaque oxide]

i don’t think they have money to host this model to the mass lmao

probably why they would try distillation

[dense halo]

makes sense to me i guess

one thing i wonder and this seem for all ai... but the context or token you can enter is always get bigger, fatter, more pages or data or whatever

can enter a book on claude i think... but why is output so small? like 2 page on all ai

[sleek goblet]

anyone else getting the "invalid request" from the macos desktop app?

[dense halo]

on purpose?

[boreal timber]

Understood. Unfortunately it stop on basically every chat

[cedar glade]

Output is more expensive than input.

[tight island]

Does OAI have a huge model behind closed doors?, of course they do, there’s no need for proof, it’d be extremely weird if they didn’t

[dense halo]

does that mean it is possible though? just limit for cost?

[wanton lichen]

I'm 100% aware that the new voice model isn't out yet, but they have definitely made some new updates to the current model.

It doesn't constantly listen and seems to have some awareness about when I'm just muttering and when it should be listening

The way it sits there and waits when it doesn't detect anything substantial is so slick

[urban cipher]

They took my access overnight lol

[cinder bison]

i got access to the mac app can i upload screenshots to it yet?

[wanton lichen]

Yes

[cinder bison]

when in voice mode i drag and drop onto the app and it uploads it to the chat but doesnt send it so i have to leave voice chat then send it

[wanton lichen]

Honestly doing it that way isn't even something I had tried. Let me try that real quick and see how it goes

[cinder bison]

No problem, just thought id give the feedback

[wanton lichen]

Wow I tried it and it actually just crashed the desktop app

Interesting

[cinder bison]

i didnt get that i just got it saying release to send but when releasing it does nothing but when i leave voice its in the chat box uploaded

my responses keep changing language too 😆

[wanton lichen]

Yep I'm getting the same thing. I actually hate that the voice takes over the whole window and I can't interact with it in any other way when using voice

I just had a weird lanuage change too lol

[cinder bison]

be nice if you could leave the voice chat window but still be connected or maybe have it on the right of the chat bar instead of a full take over

[wanton lichen]

I have a feeling this will be worked out eventually, having the chat window available at the same time opens up a bunch of interaction options

[cinder bison]

I think so too, Guess we wait for the updates haha

[jade patio]

there is no waiting list for the macos App right? i just have to wait for my turn as a plus user?

[young basin]

I'm assuming the current 'vision' in 4o is not the new one. I just uploaded a screenshot of a tier list and asked it to write out the list. It got it catastrophically wrong 😂

[cedar glade]

The new functionality isn't out yet.

[high minnow]

is gpt4o available for free users yet?

[cedar glade]

I think it is. But not the multimodal part. Just the text part.

So you won't be able to do any of the stuff in the demo.

It's still just a faster chat

Yepp

Good clarification.

[stuck narwhal]

“ We're opening up access to our new flagship model, GPT-4o, and features like browse, data analysis, and memory to everyone for free (with limits). “ hey ppl what does this free mean? Free user can use gpt4-o in ChatGPT?

[grizzled badge]

PLEASE implement some smooth scrolling of received text. It is so "choppy", "wait, wait, page of data, wait 10 small replies all at once, wait, wait, 2 more pages of data. Makes me 🤢 to look at.

[dusky canopy]

Hello. Do I have gpt-4o or not in my ChatGPT?

[grizzled badge]

Look and find out?

[dusky canopy]

bad answer

[grizzled badge]

No, it's the only answer

Load chatgpt.com and see if the model is there

I don't know what YOU have, only you do

[granite harness]

Is there anything in my mailbox?

[boreal timber]

lmao

[dusky canopy]

is this a reference I don't undersatnd?

[stray forum]

good answer

[granite harness]

Bad answer

[dusky canopy]

so many blocked users 🫠

[flat python]

Why I cannot access GPT-4o?

[maiden fractal]

You wait till available

[flat python]

Okay. But my friend can access it

I will wait. Thanks

[tall wind]

gpt is slow .

4 and 4o

network errors.

[tall summit]

Works fine for me

Just used it at work

[young basin]

Basically, in the top left of the chat screen you can see the model you are using. If there is 4o in the list then it's available. It isn't replacing ChatGPT 4

[dusky canopy]

top left doesn't say 4o

its this:

ChatGPT
Great for everyday tasks

but it doesn't tell me which version of ChatGPT

[misty ingot]

Hi, I got a question i can't seem the find the answer about it when looking on the internet not to deeply, between gpt4 and gpt4o, should I just use the newest one for complex tasks (reasonning stuff) or gpt4 is more suited for that or there is no diff ?

[marsh oyster]

hi! in the presentation video of the new gpt4 you can see how people have activated the camera on the app. where can i do this in the app, are there several apps?

[tall wind]

How to post images screen shots, or links?

[misty ingot]

drag and drop or ctrl c ctrl v, u need gpt4

[tall wind]

oh i mean, into discord

like a chart

[tall summit]

If you are on the web version it tells you

So you are probably using the app

[dry knoll]

Has anyone without the plus account received GPT4o? I am localized in europe but still haven't received anything

[cedar glade]

I thought everyone had GPT4o as a model in chatgpt. Every plus user, that is.

However, nobody has the multi-modal features.

[dusky canopy]

I am on the web version

[cedar glade]

Those are being rolled out "to a small group of trusted partners in the API in the coming weeks."

Might be a long time before they show up in ChatGPT.

😦

[dusky canopy]

[cedar glade]

This won't be available in some time, it seems.

[lavish hemlock]

Is ChatGPT-4 intentionally slow with the launch of 4o or something? lol

[marsh oyster]

ok funny that it is shown in all videos and is advertised but you can't use it, too bad...

[keen meadow]

It's super fast for me. Not sure I've noticed any slow-downs actually, and I use ChatGPT Plus a lot. Maybe capacity is divided up into regions and Africa/MEA/EMEA/whatever tends to be very fast. I've heard from friends abroad about slow responses so it's not just you.

[cedar glade]

I think they may have rushed the demo to be before google, but they weren't ready to release. Too bad, yes.

[river brook]

consider i have 2 friends, both free versions, why does one have access to gpt4o and one not?

[maiden fractal]

Cause it's rolling out

[undone saffron]

Now that I have access to 4o, is it possible to manually switch from 4o back to 3.5 if I don't want to spend the limit for my simpler questions?

I only have one choice which is ChatGPT, it says model-auto?

[tall summit]

I did.

[sacred meadow]

Out of interest, what region are you that you have access now?

[tall summit]

Weird

[undone saffron]

Asked for a friend but we're in Sweden

He has free account

[sacred meadow]

Ok thx!

[fluid jackal]

Man i thought gpt-4o would fix the slow and unresponsive chat, but nope 😄

If your convo is more than 5-ish questions it just breaks

[eternal rock]

This is a Safe for Work Discord Server

The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.

Posting NSFW content will result in a permanent ban.

[tall summit]

Guess you’re just unlucky. Works fine for me. Are you located in Europe?

[rocky wing]

Does 4o change the quality of output from the DALL-E image generator side of things at all?

[dry knoll]

where are u located?

[fluid jackal]

Yeeep, a workaround is basically waiting the amount of time you think the response takes, then reload the whole page
(and yes I've tried several devices, browsers, virtualbox etc.)

[tall summit]

I am from the Netherlands

[tall summit]

I see

[inland heron]

I'm in Europe and 4o works fine. Even outputs long lines of code without issues. I haven't had an error in 2 days, and even then happened only because of (probably) large influx of users due to announcement and that live video, and just about 2-3 errors at most. And before that, no issues as well for weeks. After that, also no issues.

The only thing I noticed, is that with complex functions, GPT tends to repeat things or doesn't try hard. But when I scold it, then it does great 😄

[tall summit]

Nice nice!

[inland heron]

really strange to see that, and I don't want to sound angry in GPT chat, but it does work for some reason. I don't know how to supplement this differently

[tall summit]

Good to hear

[tall summit]

Didn’t sound angry at all hahaha

[inland heron]

no not here, with GPT 😄

[tall summit]

Yeah

I also say “please” sometimes when entering a prompt 🤣

[solid bronze]

I tell it what to do as a command not a option lol

[inland heron]

I do, too, and it's best to stay calm and be thankful as well, but once you start getting the same result after about 10-15 iterations of the exact same test or it starts going in circles, then you have to try something different. I get stuck as GPT becomes redundant, so I either have to tell it 'off' and hope for a better output, or start a new conversation

[inland heron]

That's fine, but if you do iterations of testing the same code for like 20-30 times, it doesn't really matter if it's a command or not

it just becomes redundant, the conversation suffers, and I don't know how to get it to stop looping the same code over and over again

unless I 'push buttons'. I tried the trick of 'I'll give you a $100 tip' but doesn't work 😄

[nova wind]

Why is there 2 new chat buttons when you expand the side-panel? 💀

It feels off, + the small one adds ?model=text-davinci-002-render-sha to the end of the url.

[tall summit]

Indeed!

[lavish thicket]

Hey friends,
Where can I find the MacOS app? I had a link in a new chat today and I closed it and cannot find it since

[rancid pier]

I've seen a couple of people reporting that ChatGPT knows things about them they wouldn't expect - even when they don't have the memory feature. One person it knew what state they lived in, and another person it knew what companies they worked for.

I can't replicate anything similar myself. Has anyone come across this, or know what could cause it? Info in training data that ChatGPT has somehow matched to the user?

[fallow bear]

I got the invite to try ChatGPT desktop but when I launch it, it still says that I don't have access. Anyone else got this issue?

[solid bronze]

Just had an update on my mobile app for chat gpt but still no vision yet lol

[burnt hare]

it will be a while before they drop vision

[solid bronze]

Conversation seems to be better less robotic

Less terminator lol

But to interrupt I still have to tap the screen

In the demo they were just able to talk

[elfin birch]

The new voice mode is rolling out over the next few weeks

[solid bronze]

Im excited about that

People already think I’m crazy at my desk speaking to the ai lol

[inland heron]

Nope. But then again, I don't usually ask it to tell stuff about me - I just work on solving issues and building things

[dry quest]

yes, I am

[rancid pier]

Can't post a link here, but the community forum has a discussion titled "Chat GPT Desktop App for Mac"

[quasi adder]

When will GPT 4o come out for free users?

[deep flax]

it is already out

gradual release tho, just wait a few days and you will get it soon

[feral zinc]

its a rollout so depending of good luck some even got it on monday and some will get it in next few days or weeks

[rancid pier]

Yeah, one example was they asked how to make a job opening post engaging, a general question without giving any details, and ChatGPT gave some suggestions and a draft of a post that included the correct state.

The other example was someone asking for general advice on how to answer the question 'what achievements are you most proud of' and chatgpt cmae back with suggestions that included things like 'perhpas something you achieved while working at [actual company they worked for].'

[gaunt magnet]

why does using gpt4o count up to response limit for gpt4?

[dry knoll]

Todays status on the GPT-4o rollout:

• GPT-4o (for text) available to most paid users
• GPT-4o (for text) free version started rolling out today
• Image Generation today is still DALL-E 3
• Image Upload is the new GPT-4o
• Improved Web Browsing and Code interpreter are available today
• GPTs still use GPT4
• Voice input remains the old Whisper
• Voice output remains the old tts-1
• Mac App not available yet

[feral zinc]

dont spam and that belongs to #ai-discussions

[zealous socket]

@sleek aspen

[sleek aspen]

thanks for the ping!

[summer pagoda]

Thanks for the O version. It is faster. Lets hope it still gives good answers haha

[lavish thicket]

thank you

[eternal rock]

Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.

Please report any incident by sending a DM to @languid valley immediately.

[buoyant zinc]

The Mac App just became available.

[muted jolt]

For you? Not for all plus users yet.

Typical OpenAI to overpromise and under deliver.

Still not working for me.

[buoyant zinc]

I opened my desktop ChatGPT and there was a pop-up indicating the download for it. Working 100% now. Pretty cool.

[scarlet cave]

can i send chatgpt 4o an image link? so he can see it?

[muted jolt]

Slow roll out... They probably should have said starting the roll out slowly to plus users first. What a let down....

[zinc tulip]

Yes

[scarlet cave]

w

[zinc tulip]

The vision function is more impressive than anything else

I'm already making bots with voiceovers with different people chatting like chatgpt4-o

That is wack...whats the purpose of being OpenAI if you are restricted....

All i said was a name smh...

[burnt hare]

this is the direct link to a 3.5 chat https://chatgpt.com/?model=text-davinci-002-render-sha

[spice cypress]

I wish I can use the new voice image feature

I’m Plus user

[honest jay]

Hey there! It'll be rolling out in the coming weeks -- I'm excited too! https://openai.com/index/hello-gpt-4o/

We'll roll out a new version of Voice Mode with GPT-4o in alpha within ChatGPT Plus in the coming weeks.

[spice cypress]

Can’t wait!! 🥹🥹🥹.

[primal shard]

Does any free user got gpt 4o in europe ?

[maiden fractal]

Yes

[lean moth]

the current 4o model that ive used within the app does not seems to be as natural sounding as the demo'd version from openai on youtube, guessing that will come in a few weeks with an update

[novel parcel]

What is this

Error occured. The server had an error processing your request. Sorry about that! You can retry your request, or contact us through our help center at help.openai.com if you keep seeing this error.

[honest jay]

Yep, exactly! https://openai.com/index/hello-gpt-4o/

We'll roll out a new version of Voice Mode with GPT-4o in alpha within ChatGPT Plus in the coming weeks.

[solid bronze]

Ready for this to roll out

[gusty summit]

What about the new camera/vision thing

Will it roll out at the same time?

[marsh oyster]

its actually not work with the camera

[gusty summit]

WDYM?

[marsh oyster]

currently it only works with the voice, the camera cannot be activated live in the app

[feral zinc]

new version is not out yet

[worthy dragon]

rolling out slowing, i got it today

[primal shard]

I don't have it

[native vortex]

sucks that 4o gets 50% right text generation

better than the pretty much never of 4

[solid bronze]

Windows doesn’t have a desktop app yet right it only Mac ?

[stark wadi]

It couldnt download because its not compatable with the latest version of macos

the download file has a slash through it. says my computer (which has the latest macos) isnt compatible.

i have a mac. yes, mac only.

[inner walrus]

Mmmm. Unfortunately, I don't have the new version yet. Do I need a Pro version first?

[honest crown]

hey chatgpt fellas!

[visual bloom]

There is no pro version. Only free, plus, team, and enterprise versions.

[vivid thicket]

is there supposed to be a desktop update that lets gpt4o view your screen in realtime for live assisting?

not sure if i heard that right from the livestream

[urban cipher]

Anyone else having the issue that a custom GPT performs worse after specific instructions than before? I gave it instructions how to explain things and it got some easy problems wrong, while the normal GPT4o got it right

[sturdy cave]

do free users have access to the desktop app?

[feral zinc]

not for now. "We're rolling out the macOS app to Plus users starting today, and we will make it more broadly available in the coming weeks. We also plan to launch a Windows version later this year."

[knotty scaffold]

I got the prompt to download the macOS app, but when I open it and login it says I don't have access

anyone else have that issue?

[winged moth]

Can you share an example? I have not noticed this

[urban cipher]

I have wrote the instructions in german so you probably cant do anything with them. I basically said that the GPT should explain things without the use of complicated calculations and rather use a theoretical and logical approach. It completely ignores it and calculates everything. If i point out that it shouldnt use calculations it answers the way i want

But only if i explicitly tell it to again

[shut palm]

When free gpt 4o

[neat storm]

they rushed to deliver GPT-4o without all the cool features to beat Google to the punch

[lucid stump]

Hey!

I wonder if web search will become available for us free users eventually...

[honest jay]

Hey! When 4o rolls out to your free account, you'll indeed be able to use its web browsing feature: https://openai.com/index/gpt-4o-and-more-tools-to-chatgpt-free/

When using GPT-4o, ChatGPT Free users will now have access to features such as:
** · **Get responses from both the model and the web

[lucid stump]

nice thx!

[stark wadi]

Thw download app for macOS isnt working

[grim pumice]

Yes, same for me.

[grim pumice]

Is your Mac Intel? The app is Apple Silicon only. Typically an app with a strikethrough like that means it can't run due to processor architecture incompatibility.

[stark wadi]

yes you are correct. thank you.

[regal portal]

apparently there was a small update to the iPhone app yesterday, but I don’t see anything different in neither the user interface nor interaction with GPT4o

[muted jolt]

Pathetic lol

The pathetic race to beat google to the punch and deliver a quarter of the expected features after that demo with the "in coming weeks" statement is absolutely trash.

[maiden fractal]

Gpto have higher limit for plus and much faster

[opaque oxide]

Hey guys, will voice mode be available for free users or only the text model?

[misty vale]

No Voice Mode for free users, at least at first.

[opaque oxide]

damn, that’s sad.

[regal portal]

if you want it that bad, just pay the 20 bucks 🙂

after it becomes available, that is,

[inland heron]

anyone experiencing low limits on GPT4o?

[maiden fractal]

For paid or free ?

[inland heron]

Plus

[elder vector]

When will the new voice experience of gpt 4o be avalible for Plus user?

[maiden fractal]

After how many prompt ?

[inland heron]

Is it hard to say 'No'? 😄

because that's what you're about to say

I've never seen you say anything bad or slightly hint at low limits or bad experience with errors here