Skip to content
Open specification / 1.0

Define an LQM so systems can interoperate.

The Large Quantitative Model Specification 1.0 defines the minimum identity, contract, evaluation, evidence, governance and operation fields required to describe a governed quantitative model system consistently.

01IdentityKey · version · implementation hash
02ContractInputs · outputs · parameters
03EvaluationBenchmarks · metrics · guards
04OperationEvidence · governance · deployment

Normative minimum

An LQM manifest SHOULD declare a stable model key and version, the quantitative task, executable input/output schemas, runtime information and an implementation identity. A production-grade implementation SHOULD also expose benchmark references, execution-evidence capability, governance state and deployment state.

Architecture-neutral by design

The specification does not require one model architecture. Statistical models, econometric models, optimization systems, simulation engines, tree models and neural models can all satisfy the contract when they operate as quantitative authority.

Open resources

LargeQuant publishes machine-readable schemas so external systems can adopt the terminology without adopting the LargeQuant runtime.