Evaluators that catch failures automatically.
Patronus runs automated evaluation over LLM outputs — hallucination detection\, relevance\, safety — with purpose-built evaluator models rather than asking a general model to grade itself.
Purpose-built evaluators matter: a general model grading its own output shares its blind spots. API pricing is published per thousand calls — $10 for small evaluators\, $20 for large — with $10 in free credits and a free Developer tier to start.
Quick Information
Platform
Web
Pricing
Free + Paid (from $25/mo)
API
Available
Category
Analytics
Pros and Cons
Pros
- Purpose-built evaluator models
- Free Developer tier with no card
- $10 free credits to start
- Published per-call API pricing
- On-prem and VPC on Enterprise
Cons
- Free tier keeps only 2 weeks of history
- Page limits on the Base plan
- Evaluator calls billed separately
- Enterprise pricing not published