We (Kim, Muhn and Nikolaev) examine to what extent general-purpose large language models (LLM) such as GPT4 can make informed financial decisions based on mostly numerical financial data. We provide ...