LLama Guard 4

Llama Guard 4 is a natively multimodal safety classifier with 12 billion parameters trained jointly on text and multiple images. It is a dense architecture pruned from the Llama 4 Scout pre-trained model and fine-tuned for content safety classification. Similar to previous versions, it can be used to classify content in both LLM inputs (prompt classification) and in LLM responses (response classification). Llama Guard 4 itself acts as an LLM: it generates text in its output that indicates whether a given prompt or response is safe or unsafe, and if unsafe, it also lists the content categories violated. Llama Guard 4 was aligned to safeguard against the standardized MLCommons hazards taxonomy and designed to support multimodal Llama 4 capabilities within a single safety classifier. Specifically, it combines the capabilities of the previous Llama Guard 3-8B and Llama Guard 3-11B-vision models by supporting English and multilingual text prompts (on the languages supported by Llama Guard 3) as well as mixed text-and-image prompts for image understanding. Unlike Llama Guard 3-11B-vision, Llama Guard 4 now supports safety classification when multiple images are given in the prompt as input. Llama Guard 4 is also integrated into the Llama Moderations API for text and images.