Include model modalities on hover modal

fs1010

I think the design of this is good, but including icons to indicate modalities other than text would be very useful as well. Such as an image icon for image input support, or a document icon for PDF support. Also, maybe include a warning in the OpenAI models that requests sent to OpenAI will be logged.

Example:

This will make it easier for users to decide which model to choose based on what they’re doing and what features they require.

Thibaultmol

So technically all models currently listed work with pdf's because currently they all go through the same parser to turn into a text file for the model.
(there are plans to make gemini and such who can process pdf's natively, do that. but that's not the case yet).

But def for image input that's a fair feature to have yeah

Thibaultmol

Update on this btw, there is now a text that says if the model doesn't support images natively (as most models these days do support images)

example^