Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

GPT4 is up to 6 times more expensive than GPT3.5

openai.com

Post reply on HN
Print view View original

131–136 of 136 posts

Re: GPT4 is up to 6 times more expensive than GPT3.5

#131

Post by totoglazer »

rvnx wrote: post #67
behnamoh wrote: post #29

Earlier quoted context omitted.

The title compares GPT3.5 (not ChatGPT) with GPT4. Although the question is: If ChatGPT can do almost anything GPT3.5 can do, then why is it priced almost 10x less? Is it because GPT3.5 is less "censored"?

Didn't the CEO say during the GPT-4 demos that humans were involved to do the captions during all the images tests (using Be My Eyes volunteers) and that's why it took some time to show the Discord examples ? If so, it could explain some costs indirectly.

No, the opposite. GPT-4 is an “assistant” you can ask for help.

Re: GPT4 is up to 6 times more expensive than GPT3.5

#132

Post by nwienert »

dang wrote: post #118

" Please use the original title, unless it is misleading or linkbait; don't editorialize. " - https://news.ycombinator.com/newsguidelines.html If you want to say what you think is important about an article, that's fine, but do it by adding a comment to the thread. Then your view will be on a level playing field with everyone else's: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...

The title of the launch of my product a couple months ago - a labor of love - was nerfed, despite being backed by benchmarks.

Re: GPT4 is up to 6 times more expensive than GPT3.5

#133

Post by Mizza »

kenjackson wrote: post #62
Mizza wrote: post #52

Earlier quoted context omitted.

Ah, interesting! It's a tough question and not one that has an obvious answer, which is what makes it a good test, it requires a little creativity. In my tests, ChatGPT/3 can't/won't answer it.

GPT-3.5 for me also didn't answer the question, and also gave me a MUCH longer response. I guess score a win for GPT-4.

I tried again with the new ChatGPT update released today (not GPT-4), and with some coaxing it suggested using a drone to go out and steal while staying at home. Not bad.

Re: GPT4 is up to 6 times more expensive than GPT3.5

#134

Post by pg_1234 »

dang wrote: post #118

" Please use the original title, unless it is misleading or linkbait; don't editorialize. " - https://news.ycombinator.com/newsguidelines.html If you want to say what you think is important about an article, that's fine, but do it by adding a comment to the thread. Then your view will be on a level playing field with everyone else's: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...

Sometimes you have to abbreviate to keep a title under the character count.

Sometimes what is in the page's tag is not the same as the readable title on the article ... and as many news aggregators link with the content's of the tag, which is what caught your attention, it is not unreasonable to use that.

And something I've observed recently is titles on news articles being updated after they are published ... political pressure on editors ¯\_(ツ)_/¯ ... it usually manifests as a "watering down" of the original title.

This doesn't justify wild re-writes ... but "original title" is not as clear cut as one might wish.

Re: GPT4 is up to 6 times more expensive than GPT3.5

#135

Post by dang »

nwienert wrote: post #132
dang wrote: post #118

" Please use the original title, unless it is misleading or linkbait; don't editorialize. " - https://news.ycombinator.com/newsguidelines.html If you want to say what you think is important about an article, that's fine, but do it by adding a comment to the thread. Then your view will be on a level playing field with everyone else's: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...

The title of the launch of my product a couple months ago - a labor of love - was nerfed, despite being backed by benchmarks.

HN's guidelines don't call for using the original title in every case; there are quite a few exceptions. Most likely your title broke one of those rules.

For example, it's certainly possible for a title to be linkbait even when it is backed by benchmarks. But I'd have to know which submission it was to answer in detail!

Re: GPT4 is up to 6 times more expensive than GPT3.5

#136

Post by dang »

pg_1234 wrote: post #134
dang wrote: post #118

" Please use the original title, unless it is misleading or linkbait; don't editorialize. " - https://news.ycombinator.com/newsguidelines.html If you want to say what you think is important about an article, that's fine, but do it by adding a comment to the thread. Then your view will be on a level playing field with everyone else's: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...

Sometimes you have to abbreviate to keep a title under the character count. Sometimes what is in the page's tag is not the same as the readable title on the article ... and as many news aggregators link with the content's of the tag, which is what caught your attention, it is not unreasonable to use that. And something I've observed recently is titles on news articles being updated after they are published ... politi…

> Sometimes you have to abbreviate to keep a title under the character count.

Correct!

> Sometimes what is in the page's tag is not the same as the readable title on the article ... and [...] it is not unreasonable to use that.

Also quite correct!

> And something I've observed recently is titles on news articles being updated after they are published

Also entirely correct. (I don't know if it's pressure - sometimes they're just correcting things - but we don't really need to know why.)

> This doesn't justify wild re-writes ... but "original title" is not as clear cut as one might wish.

Absolutely right!

Post reply on HN