PNG vs other file formats for GPT Vision

API
1

Hi all,

does anyone know if any one particular image format (e.g. PNG) works better than others with the vision model?

Thanks,
Raivat

2

I don’t think it really matters, because the image will be tiled and tokenized anyways, and possibly resized. It’s good to keep this in mind: https://platform.openai.com/docs/guides/vision/low-or-high-fidelity-image-understanding

3

Ahh okay, thanks so much!