OpenAI has introduced a new multimodal moderation model, “omni-moderation-latest,” built on GPT-4o. It supports both text and image moderation, enhancing accuracy, especially for non-English languages. The model classifies content across multiple harm categories, providing developers with more granular control over moderation. It’s free for developers via the Moderation API.
Source: Upgrading the Moderation API with our new multimodal moderation model
Leave a Reply