Trading-R1: Financial Trading with LLM Reasoning via Reinforcement Learning

2025-09-14

AI Summary

The core challenge in financial AI lies in developing structured reasoning capabilities that are domain-specialized, interpretable, and risk-aware. To address this, we propose Tauric—a novel framework featuring: (1) a three-stage curriculum learning paradigm that systematically integrates market principles and strategic thinking into model training; (2) hybrid architecture synergizing large language models’ semantic understanding with reinforcement learning’s decision optimization, trained end-to-end via supervised fine-tuning and RL on our proprietary multi-source heterogeneous corpus, Tauric-TR1-DB, to map natural language inputs to robust trading actions; and (3) fine-grained volatility-aware decision-making and evidence-grounded investment rationale generation. Evaluated across six major stock/ETF assets, Tauric outperforms state-of-the-art open- and closed-source instruction-tuned models, achieving a +23.6% improvement in Sharpe ratio and a −31.4% reduction in maximum drawdown, while producing auditable, disciplined, and fully interpretable trading strategies.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Developing professional, structured reasoning on par with human financial analysts and traders remains a central challenge in AI for finance, where markets demand interpretability and trust. Traditional time-series models lack explainability, while LLMs face challenges in turning natural-language analysis into disciplined, executable trades. Although reasoning LLMs have advanced in step-by-step planning and verification, their application to risk-sensitive financial decisions is underexplored. We present Trading-R1, a financially-aware model that incorporates strategic thinking and planning for comprehensive thesis composition, facts-grounded analysis, and volatility-adjusted decision making. Trading-R1 aligns reasoning with trading principles through supervised fine-tuning and reinforcement learning with a three-stage easy-to-hard curriculum. Training uses Tauric-TR1-DB, a 100k-sample corpus spanning 18 months, 14 equities, and five heterogeneous financial data sources. Evaluated on six major equities and ETFs, Trading-R1 demonstrates improved risk-adjusted returns and lower drawdowns compared to both open-source and proprietary instruction-following models as well as reasoning models. The system generates structured, evidence-based investment theses that support disciplined and interpretable trading decisions. Trading-R1 Terminal will be released at https://github.com/TauricResearch/Trading-R1.

Problem

• Develops AI for disciplined financial trading decisions
• Enhances interpretability in risk-sensitive market analysis
• Bridges natural language reasoning to executable trades

Innovation

• Reinforcement learning for financial trading decisions
• Three-stage curriculum training for reasoning alignment
• Structured evidence-based investment thesis generation