Finance Language Model Evaluation (FLaME)

2025-06-18

AI Summary

Existing financial NLP evaluation frameworks suffer from methodological limitations, leading to systematic underestimation of large language models’ (LLMs) performance on domain-specific financial tasks. To address this, we introduce FLaME—the first comprehensive, finance-oriented LLM benchmark, encompassing 20 core FinNLP tasks and 23 foundational models. FLaME pioneers a “reasoning-enhanced” comparative evaluation paradigm, integrating domain-adapted prompt engineering, chain-of-thought assessment, and fine-grained error analysis. Empirical results demonstrate that mainstream LLMs possess substantially greater financial reasoning capability than previously reported. All benchmark data, evaluation code, model outputs, and the full benchmark suite are fully open-sourced to ensure reproducibility and foster rigorous, transparent financial AI research.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Computer Vision: Large Vision Models

Application Category

• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Language Models (LMs) have demonstrated impressive capabilities with core Natural Language Processing (NLP) tasks. The effectiveness of LMs for highly specialized knowledge-intensive tasks in finance remains difficult to assess due to major gaps in the methodologies of existing evaluation frameworks, which have caused an erroneous belief in a far lower bound of LMs'performance on common Finance NLP (FinNLP) tasks. To demonstrate the potential of LMs for these FinNLP tasks, we present the first holistic benchmarking suite for Financial Language Model Evaluation (FLaME). We are the first research paper to comprehensively study LMs against'reasoning-reinforced'LMs, with an empirical study of 23 foundation LMs over 20 core NLP tasks in finance. We open-source our framework software along with all data and results.

Problem

• Assessing LM effectiveness for specialized finance tasks
• Addressing gaps in financial NLP evaluation methodologies
• Benchmarking LMs on reasoning-reinforced finance NLP tasks

Innovation

• Developed holistic benchmarking suite FLaME
• Studied reasoning-reinforced LMs comprehensively
• Open-sourced framework with data