MASFIN: A Multi-Agent System for Decomposed Financial Reasoning and Forecasting

TL;DR

MASFIN is a multi-agent AI framework that combines large language models with financial data and bias mitigation to improve short-term stock portfolio performance, demonstrating promising results over traditional benchmarks.

Contribution

Introduces MASFIN, a modular multi-agent system integrating LLMs with financial metrics and bias protocols for transparent, reproducible financial forecasting.

Findings

Achieved 7.33% weekly return over eight weeks.

Outperformed major stock indices in six of eight weeks.

Demonstrated the effectiveness of bias-aware AI in finance.

Abstract

Recent advances in large language models (LLMs) are transforming data-intensive domains, with finance representing a high-stakes environment where transparent and reproducible analysis of heterogeneous signals is essential. Traditional quantitative methods remain vulnerable to survivorship bias, while many AI-driven approaches struggle with signal integration, reproducibility, and computational efficiency. We introduce MASFIN, a modular multi-agent framework that integrates LLMs with structured financial metrics and unstructured news, while embedding explicit bias-mitigation protocols. The system leverages GPT-4.1-nano for reproducability and cost-efficient inference and generates weekly portfolios of 15-30 equities with allocation weights optimized for short-term performance. In an eight-week evaluation, MASFIN delivered a 7.33% cumulative return, outperforming the S&P 500, NASDAQ-100, and Dow Jones benchmarks in six of eight weeks, albeit with higher volatility. These findings demonstrate the promise of bias-aware, generative AI frameworks for financial forecasting and highlight opportunities for modular multi-agent design to advance practical, transparent, and reproducible approaches in quantitative finance.