Llama Guard 4

A 12B parameter Meta safeguard model that reads text and images together and labels content safe or unsafe with the violated categories out of fourteen. Checks both the user prompt and the model reply; a drop-in replacement for Llama Guard 3.

оптимизирована под английский текст; классификацию одного изображения без текста и проверку сгенерированных картинок не поддерживает

Understands

text, images

Produces

text

Contents verified with the vendor: 2026-08-18

Included in subscriptions

No catalogue entry marks this model as part of a plan yet. It will appear here once the plan contents are confirmed with the vendor.