⚡️ FlashAttention-4: up to 1.3× faster than cuDNN on NVIDIA Blackwell →

Introducing Together AI's new look →

🔎 ATLAS: runtime-learning accelerators delivering up to 4x faster LLM inference →

⚡ Together GPU Clusters: self-service NVIDIA GPUs, now generally available →

📦 Batch Inference API: Process billions of tokens at 50% lower cost for most models →

🪛 Fine-Tuning Platform Upgrades: Larger Models, Longer Contexts →

Models / Meta
Moderation

Llama Guard 2 8B

8B Llama 3-based safeguard model for classifying LLM inputs and outputs, detecting unsafe content and policy violations.

Read our tutorial
read docs

About model

Llama Guard 2 8B detects and flags potentially harmful or sensitive content, serving as a robust tool for developers and content moderators seeking to ensure online safety and compliance. Its key strength lies in accurately identifying nuanced and context-dependent threats. It is designed for professionals managing online platforms and communities.

To use this moderation model, please follow the instructions from our blog post.

    Related models
    Moderation
    VirtueGuard Text Lite
    • Input/1M
      $0.20
    • Context
      131K
    No items found.
    Moderation
    New
    Moderation
    Llama Guard 4 12B
    • Input/1M
      $0.20
    • Output/1M
      $0.20
    • Parameters
      12B
    No items found.
    Moderation
    Moderation
    Llama Guard (7B)
    • Parameters
      7B
    No items found.
    Moderation
    Model specifications
    Model data
    • Model provider
      Meta
    • Type
      Moderation
    • Main use cases
      Small & Fast
      Moderation
    • Deployment
      Monthly Reserved
    • Parameters
      8B
    • Context length
      8K
    • Input price

      $0.20 / 1M tokens

    • Output price

      $0.20 / 1M tokens

    • Input modalities
      Text
    • Output modalities
      Text
    • Released
      April 18, 2024
    • Last updated
      February 24, 2026
    • External link
      Provider docs
    • Category
      Moderation
    Quickstart docs
    Deploy model