I noticed AI Gateway does not provide access to SOTA OCR models, or any OCR models for that matter. Would love to see all OCR models there.

We have tons of volume and perfect use case for OCR models, and I'm sure more people do too. Pretty please.