New benchmark tests AI agents on real financial reasoning tasks

Original: Hedge-Bench: Benchmarking Agents on Hard, Realistic Tasks Pertaining to Financial Reasoning

Source: arxiv.org ↗