FinRL Contests: Data‐Driven Financial Reinforcement Learning Agents for Stock and Crypto Trading

ABSTRACT Financial reinforcement learning (FinRL) is now a practical paradigm for financial engineering. However, applying RL strategies to real‐world trading tasks remains a challenge for individuals, as it is error‐prone and engineering‐heavy. The stock and crypto trading tasks pose unique challenges due to the non‐stationarity of financial data, low signal‐to‐noise ratios and various market frictions. Although numerous FinRL methods have been developed for stock and crypto trading tasks, the lack of standardised task definitions, real‐time high‐quality datasets, close‐to‐real market environments and robust baselines has hindered consistent reproduction in both the open‐source community and the FinTech industry. To bridge this gap, we organised a series of FinRL Contests from 2023 to 2025, covering a diverse range of financial tasks such as stock trading, order execution, crypto trading, and the use of large language model (LLM)‐engineered signals. These contests attracted 230+ participants from 100+ institutions in 20+ countries. To encourage participation, we provided starter kits that feature GPU‐optimised parallel market environments, ensemble learning methods, and comprehensive instructions. In this paper, we summarise these benchmarking efforts, detailing trading task formulations, data curation pipelines, environment implementations, evaluation protocols, participant performance and organisational insights. It guides our follow‐up FinRL Contests and also provides a reference pipeline for FinAI contests alike.