Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
Prompt-Guardhuggingface.co · Low evidencePrompt-Guard-86M is a small Meta Llama 3.1-based language model hosted on Hugging Face, intended for text classification and prompt-safety tasks. It is released under the Llama 3.1 Community License, permitting use, modification, and redistribution with attribution and requirements to label outputs and include 'Built with Llama' in distributed materials. Access requires agreeing to share contact information, and the license includes commercial terms (notably a 700 million monthly active users threshold) and a standard warranty disclaimer.
The market map
LLM Firewall and Guardrail Platforms →Where AI ranks Prompt-Guard