Build and own a Treasury Investment Specialist Model.
How should excess cash and treasury assets be allocated across investment instruments to balance yield, liquidity, and risk?
Overview
Optimizes short-term capital allocation across instruments to improve yield without breaching liquidity or risk constraints.
Who owns this decision
Treasurer, VP Treasury
Why owned AI may fit
Treasury-investment-engine is the paired allocation-decision engine to treasury-risk-ai-engine: treasury-risk-ai-engine monitors and forecasts FX, rate, counterparty, and liquidity risk exposure, while this engine makes the investment allocation decision that must operate within those risk constraints.
Model candidates to evaluate
- Qwen3.5-27B (Apache 2.0): fine-tunable reasoning backbone for allocation optimization under multiple constraints
- GLM-5.2 (MIT): long-context reasoning across full portfolio holdings, policy documents, and market data
- NuExtract 3 (Apache 2.0): structured extraction of instrument terms and policy constraints from treasury documents
Relevant benchmark starting points
- Finben
These are candidates, not performance claims. Run them against your evaluation set.


