102,558 Wikipedia Articles Preserved

Large quantitative models (LQMs) are a category of generative AI models with numerical reasoning and statistical modeling capabilities that can generate synthetic data.

The conceptual foundations associated with large quantitative models predate the term itself and are rooted in statistical modeling, quantum computing and domain-specific machine learning. Techniques for structured data analysis such as hot-deck imputation and k-nearest-neighbor imputation, established early approaches for modeling and reconstructing incomplete numerical datasets.

In the late 2010s and early 2020s, improvements in transformer architectures led to the creation of large language models designed for specific fields, especially finance and other areas that use quantitative methods. Early examples included FinBERT, a version of BERT adapted for finance, and later, large systems like BloombergGPT, which showed the benefits of training language models on financial data. The term large quantitative models was first used in a December 2023 whitepaper from FinanceGPT Labs. This document described these models as a type of generative AI system created for financial forecasting and numerical modeling.

During 2024, the term began to appear more broadly in industry and technology commentary. Publications such as Forbes described LQMs as part of a broader shift toward domain-specific AI systems designed for scientific and quantitative applications.

At the same time, companies such as SandboxAQ began promoting LQMs as a distinct category of enterprise AI, emphasizing their application in chemistry, materials science, finance and other quantitative domains.

In 2025 the term had gained wider visibility in industry and policy discussions. Organizations such as the World Economic Forum referenced large quantitative models in the context of scientific discovery and drug development.