Omni-moderation is the most capable moderation model, accepting both text and images as input. It can identify potentially harmful content across multiple categories.
Performance: High Speed: Medium Input: Text, image Output: Text