Looks like the latest UI updates for the pricing pages (1, 2) have removed the ability to calculate the cost of processing images in chat completions. Is there a separate page we should be accessing to see image processing costs for models that support vision?
Hi @tncintra ! Thanks for flagging this!
Not showing on both web and mobile?
Hi again @tncintra ! Looks like it was moved, you can see it here now:
i.e. this page, under “How is pricing calculated for images?”
Awesome lol I’m blind, thanks!
There’s now two pricing pages in operation:
https://openai.com/api/pricing/
https://platform.openai.com/docs/pricing
The URL for the second new one is in the platform site, but being in /docs/ suggests it is not some per-organization pricing possibility. The second is the one with no image token calculator or formulas.
Neither makes the clear disclosure needed that images on gpt-4o-mini are actually twice the input expense as gpt-4o.
Thanks. Even better catch.
I’ve passed it on.
Also, the language “all snapshots” in that new pricing page suggests that AI models are immutable and reliable.
Rather, versioned models present or past get stealth updates all the time that affect the quality and ability to perform the API application one has designed.
“Model names by release date” would be less deceptive and align expectations.
The presentation is odd. You have the pointer/alias of “currently recommended”. It already shows the destination. Then expanding shows more models of that class under the alias, but models to which it does not point.