FyBench

Coming soon

An upcoming benchmark evaluating how well AI systems reason through real-world financial and business tasks—from invoices and cash flow to risk, compliance and decision-making. FyBench also assesses whether models handle sensitive business information securely, follow permission boundaries, avoid unsafe or unauthorised actions, and behave reliably in high-trust environments.