Error: [404:invalid_request_error] The model: gpt-4 does not exist

[emilBeBri]

Hi so as per this announcement:

https://openai.com/blog/gpt-4-api-general-availability

everyone including me finally have access to gtp 4. but when i try:

bfp -m 'gpt-4' 'which language model are you?'

I get this:

Error: [404:invalid_request_error] The model: `gpt-4` does not exist

works fine with:

bfp -m 'gpt-3.5-turbo' 'which language model are you?'

Best,
Emil

[bakks]

Hey @emilBeBri - I'm not able to reproduce this problem, passing gpt-4 in as the model is working for me. I would try the following:

1. Create a new OpenAI API key and replace your current key in ~/.config/butterfish/butterfish.env
2. Upgrade butterfish (most recent version is v0.1.1)

I bet that (1) will resolve the problem but let me know if it doesn't.

[emilBeBri]

Hi! Ok so I made a new key and that didn't help - tried making two actually to be sure. I can see in my openai api key page that they have been used today when using

butterfish prompt 'which language model are you?' -m 'gpt-3.5-turbo'

but

butterfish prompt 'which language model are you?' -m 'gpt-4'

errors out all the same. so a new key does not do anything it seems.

running the go install to update gives me:

go install github.com/bakks/butterfish/cmd/butterfish@latest


>go: github.com/bakks/butterfish/cmd/butterfish@latest (in github.com/bakks/[email protected]):
>        The go.mod file for the module providing named packages contains one or
>        more replace directives. It must not contain directives that would cause
>        it to be interpreted differently than if it were the main module.

searching for this on google does not give any apparent solution, other than it seems to be related to the switch from go get to go install. (go get is deprecated, tried it too)

there is an issue about it here, but I'm not familiar with go so I can't really make anything out of it:

golang/go#44840

[bakks]

Ok I fixed the go install, I hadn't realized that was broken. That should now work, but I suspect probably won't fix the GPT-4 issue. Still investigating.

[bakks]

@emilBeBri If that doesn't solve the GPT-4 issue for you can you paste a screenshot of what you see at https://platform.openai.com/account/rate-limits ?

[emilBeBri]

Updated butterfish with Go just now and works great.

Hmm that's interesting. I don't see gtp-4 there. But didn't that announcement say that it was available for all paying API customers? Just read it again, and checked my settings doesn't seem like you're supposed to 'enable' the access in someway. Or maybe they're just rolling it out slow and it will be available during the coming weeks (but they didn't mention that in the announcement, either).

I'm writing a post on reddit on the chatgtp sub to hear if I'm the only one with this issue

[bakks]

@emilBeBri I talked to someone at OpenAI, apparently they don't turn on GPT-4 until you successfully pay an invoice (this is a fraud prevention measure). So if you turned on billing this month, I bet it won't enable GPT-4 until early next month after the next billing cycle. Hope that helps.

[emilBeBri]

Hi yes that's what I got out of it too.

I couldn't understand it though because I've been using butterfish on a regular basis, but only a little bit, since 3.5 is a little too simple for many purposes I find. So I looked at my billing history, and there was none!

for some reason they haven't billed me for it, like at all....! Must be a bug, or because my use has been so neglible, because now that I've created a new api key and used it for whisper, I got a small bill yesterday.

Mystery solved, thanks the help