Comment by valzam

What experience have people had with Mistral models for cyber research? are they as constrainted as Anthropic models? I use claude day2day but have the need for a model with fewer guardrails to pentest our own APIs.