Llama Guard

6.6easy start · #8 of 23
in AI Content Moderation Software

Free to practise onyes
Free trialnot on record
Well documentedyes
Runs where you worknot on record

Runs on self-hosted.

Llama Guard is ranked #8 of 23 in AI content moderation software on The Geeks Club. It runs on Self-hosted. There is a free plan.

Llama Guard plans and pricing

All plans

Llama Guard 3-8B Free 8B parameters · model weights required github.com · 4 Oct 2026

Compared on AI content moderation software

Text moderation: Yes
Image moderation: Yes
Custom policies: Yes
Deployment options: self_hosted

Best Llama Guard alternatives

See all 12

WebPurify 0 of 4 in the starter kit Lesson 17.9 Sightengine 0 of 4 in the starter kit Lesson 27.7 Google Cloud Vision API 0 of 4 in the starter kit Lesson 37.2 Hive Moderation 0 of 4 in the starter kit Lesson 47.1 Moderation API 0 of 4 in the starter kit Lesson 57.0 Besedo 0 of 4 in the starter kit Lesson 66.9

The record

Purpose: Llama Guard classifies content in LLM inputs and responses as safe or unsafe.
Model: Llama Guard 3-8B is a Llama 3.1 8B pretrained model fine-tuned for content safety classification.
Hazard categories: Llama Guard 3-8B covers 14 categories, including the 13 MLCommons hazard categories and Code Interpreter Abuse.
Languages: It supports content safety classification in English, French, German, Hindi, Italian, Portuguese, Spanish, and Thai.
Tool use: It was optimized for safety and security classification of search and code interpreter tool calls.
Customization: Llama Recipes documentation explains how to configure and customize Llama Guard 3.
Quantization: An int8 quantized version reduces checkpoint size by about 40% with very small impact on model performance, according to the model card.
Recommended use: Meta recommends deploying Llama Guard 3 alongside Llama 3.1, and says it may increase refusals of benign prompts.
Limitations: The model card says performance can be limited by pretraining data and that the model may be vulnerable to adversarial or prompt injection attacks.
Evaluation caveat: The model card notes that some categories, including defamation, intellectual property, and elections, may require factual, up-to-date knowledge for accurate evaluation.
Integration example: Meta’s Llama Cookbook demonstrates using Llama Guard as an inference safety checker for both user prompts and model outputs with Hugging Face Transformers.
Weights access: The Cookbook demo lists access to Llama Guard model weights on Hugging Face as a requirement.
Price from: Free

Where it ranks on The Geeks Club

Best AI Content Moderation Software in 2026#8 of 23

Is Llama Guard yours?

Claim it for free: prove the domain, then correct facts, plans and screenshots. An editor reviews every change.

Claim it free Promote Llama Guard

Sources

github.com/meta-llama/PurpleLlama/blob/main/Llama-· checked 4 Oct 2026
github.com/meta-llama/llama-cookbook/blob/main/get· checked 4 Oct 2026