← Back to hard AIs

Verify critical details — pricing, licensing, availability — with the model's source before business decisions. Full methodology →

Models · Meta

Llama Prompt Guard 2 22M

Model family: llama-prompt-guard

Meta's smallest prompt-injection detector — 22M parameters, sub-millisecond CPU inferenceRunning a model to get outputs — as opposed to training it. When you send a prompt to ChatGPT, that's inference. Inference is much cheaper than training per operation but adds up quickly at scale. Pricing pages almost always refer to inference costs (per million tokens, per request, etc.), not training costs., English-only. Targeted at high-throughput pipelines where the 86M variant's latency or cost is prohibitive.

Identity

Creator
Meta
Model family
llama-prompt-guard
Release date
2025-04-28

Technical specs

Parameter count
22M
Context window
512 tokens
Modalities
  • Text
Primary capabilities
  • Classification

License

License
Llama 4 Community License Agreement
Commercial use
  • Conditional

Free for commercial use unless the licensee's product has 700 million monthly active users measured at Llama 4 release date.

Terms
  • Modification
  • Redistribution
  • Attribution

Access

Openness
  • Open Weight
Access methods
  • Local Runtime Ollama
  • Weights Download Direct
  • Weights Download Hf
Cost tier
  • Self Hosted Only

Full model card →