Model Armor supports screening prompts and responses up to 65,536 tokens (262,144 characters) for prompt injection and jailbreak detection, responsible AI, and…

Model Armor includes enhanced prompt injection and jailbreak protection for Workspace data.

This capability improves detection accuracy and minimizes false positives when screening Workspace content, such as emails, documents, and files.

Template-specific exclusion rules are available in Preview.

This feature lets you configure dictionary (word and phrase lists) and regular expression rules to mitigate false-positive detections for prompt injection and jailbreak detection and responsible AI filters.