Comment Comment by acid__ Top multimodal models have about 200-1000 tokens per image, so the math works out. parent acid__ 2 years ago
