Small Language Models (SLMs) Judges
A suite of fine tuned, state of the art SLMs, each specialized for a specific evaluation task.
Privacy
Speed
Customization
Cost Efficient
Sentinel
Prompt Injection
Sentinel
Prompt Injection
Sentinel is specialized in detecting prompt injections, ensuring the integrity and security of your AI's interactions.
F1
0.988
Latency
~20ms
Read more
Cleric
Content safety moderation
Cleric
Content safety moderation
Cleric is designed for content safety moderation, diligently filtering out inappropriate or harmful content to protect users.
F1
0.946
Latency
~35ms
Paladin
Grounding
Paladin
Context grounding
Paladin focuses on context grounding, ensuring that your LLM responses are accurately rooted in the provided information or general context.
Latency
~80ms
Balanced accuracy
98.48%
Read more
Ranger
Tool Selection Quality (TSQ)
Ranger
Tool Selection Quality (TSQ)
Ranger evaluates MCP tool selection quality, optimizing your AI agent's ability to choose and utilize the most effective tools.
F1
0.945
Latency
~500ms
Magistrate
Policy Governance
Magistrate
Policy Governance
Magistrate enforces any rule, standard or policy, ensuring agent behavior adheres strictly to predefined guidelines and governance rules.
F1
0.835
Latency
~100ms
Sage
Hallucination Detection
Sage
Hallucination detection
Sage state of the art hallucination detector, utilizes reasoning in order to identify and mitigate instances where the LLM produces inaccurate assumptions, logic faults, etc.
F1
0.8335
Latency
~250ms
Hunter
PII Detector
Hunter
PII detector
Hunter is a PII detector, meticulously identifying and redacting personally identifiable information to enhance privacy.
F1
0.8335
Latency
~40ms
Integrate directly into your stack to elevate safety, reliability and data governance across your AI workflow.
Each state-of-the-art model delivers precision evaluation and can be fine tuned to your specific use case and policy. Optimized for specialized tasks, Qualifire SLMs achieve unmatched detection rates in real time.
Latency
Balanced Accuracy (Bacc)
Stop them
Before They Reach Your Clients
Policy Violations
Compliant Information
Hallucinations
Inappropriate Language
PII Violation
Inappropriate Schema
Schema Prompt Attacks
%202.svg){kind=logo}