Llama Guard 4 12B

A 12-billion-parameter content safety classifier from Meta that detects harmful content in text across multiple risk categories.

Strengths

Designed specifically for safety classification tasks, with a larger parameter count than its predecessor Llama Guard 3 8B to handle more nuanced risk detection.

Best for

Filtering user inputs and model outputs for harmful content in production systems that need reliable content moderation at scale.

Limitations

A specialized safety classifier, not a general-purpose language model; it is built to categorize risk rather than generate text or perform general reasoning tasks.