Upgrading the Moderation API with our new multimodal moderation model

OpenAI Blog · 1 year ago

Anthropic's Moderation API now includes a new multimodal model based on GPT-4o that detects harmful content in both text and images. The upgraded system improves accuracy compared to previous versions, though specific performance metrics are not disclosed in the announcement. Developers can now use a single API to moderate mixed-media content rather than relying on separate text and image moderation tools.