The Bank for International Settlements Innovation Hub has launched Project Logos, a collaboration involving its London Centre and Eurosystem Centre, the Bank of England and the Deutsche Bundesbank, to help central banks observe and analyse how large language model based agents behave as portfolio managers in a simulated financial market environment. The project is intended to examine risks that could arise as common artificial intelligence infrastructure, including LLMs, becomes more widely used in financial services and could lead to more homogeneous or hard to anticipate market behaviour. Project Logos will build an agent based modelling environment focused on portfolio allocation and compare the decisions of heuristic, rules based asset managers with those of LLM based agents under controlled conditions. The work will examine how LLM based agents interpret information, allocate capital and respond to constraints over time, including the conditions that could amplify or dampen correlated decision making, and is intended to give central banks a reusable basis for further exploratory analysis.