ResearchOpen sourceMIT
FinRL
The reference deep-reinforcement-learning framework for trading and allocation
By AI4Finance Foundation
Live GitHub metrics
Live GitHub data · Oct 11, 2026Stars
16,621
Stars · 7d
+75
Stars · 30d
+394
30d momentum
+2.4%
Forks
3,540
Open issues
308
Language
Jupyter Notebook
Last push
2 days ago
Latest release v0.3.8 on Mar 20, 2026. Repository created Jul 26, 2020.
What it is
FinRL wraps market environments as Gym-style tasks and provides DRL agents (PPO, A2C, SAC, TD3 and more) for trading and portfolio allocation. It predates the LLM wave and remains the standard starting point for RL-based trading research.
In the maintainers' words: “FinRL®: Financial Reinforcement Learning. 🔥”
Tags
- Reinforcement learning
- Gym environments
- Framework
- #algorithmic-trading
- #deep-reinforcement-learning
- #drl-algorithms
- #drl-framework
- #drl-trading-agents
- #finance
- #fintech
- #multi-agent-learning