AFAIK the costs have not been published, but assuming you'd have access to the data (which is tricky, and getting an different, equivalent dataset may involve l

AFAIK the costs have not been published, but assuming you'd have access to the data (which is tricky, and getting an different, equivalent dataset may involve lots of time/effort/legal issues/cleanup work), my estimate is that the compute for this would cost in the ballpark of $1m-$10m. But if you don't want a universal model but just a specific niche (e.g. furries) then you should be able to train a much smaller model on much less data and get something interesting much cheaper.