Open Source Reinforcement Learning Frameworks
Sort By:
Reinforcement Learning Frameworks
View 30 business solutionsBrowse free open source Reinforcement Learning Frameworks and projects below. Use the toggles on the left to filter open source Reinforcement Learning Frameworks by OS, license, language, programming language, and project status.
-
Custom VMs From 1 to 96 vCPUs With 99.95% UptimeLive migration and automatic failover keep workloads online through maintenance. One free e2-micro VM every month.
-
Veeam Data Platform v13.1 - Get Your Free TrialTry Veeam Data Platform today. Experience the unified platform that's secure by design, portable by default, and proven to recover clean, fast, and anywhere.
-
1
AirSim
A simulator for drones, cars and more, built on Unreal Engine
AirSim is an open-source, cross platform simulator for drones, cars and more vehicles, built on Unreal Engine with an experimental Unity release in the works. It supports software-in-the-loop simulation with popular flight controllers such as PX4 & ArduPilot and hardware-in-loop with PX4 for physically and visually realistic simulations. It is developed as an Unreal plugin that can simply be dropped into any Unreal environment. AirSim's development is oriented towards the goal of creating a platform for AI research to experiment with deep learning, computer vision and reinforcement learning algorithms for autonomous vehicles. For this purpose, AirSim also exposes APIs to retrieve data and control vehicles in a platform independent way. AirSim is fully enabled for multiple vehicles. This capability allows you to create multiple vehicles easily and use APIs to control them. -
2
DeepSeek R1
Open-source, high-performance AI model with advanced reasoning
DeepSeek-R1 is an open-source large language model developed by DeepSeek, designed to excel in complex reasoning tasks across domains such as mathematics, coding, and language. DeepSeek R1 offers unrestricted access for both commercial and academic use. The model employs a Mixture of Experts (MoE) architecture, comprising 671 billion total parameters with 37 billion active parameters per token, and supports a context length of up to 128,000 tokens. DeepSeek-R1's training regimen uniquely integrates large-scale reinforcement learning (RL) without relying on supervised fine-tuning, enabling the model to develop advanced reasoning capabilities. This approach has resulted in performance comparable to leading models like OpenAI's o1, while maintaining cost-efficiency. To further support the research community, DeepSeek has released distilled versions of the model based on architectures such as LLaMA and Qwen. -
3
DeepSeek-V3
Powerful AI language model (MoE) optimized for efficiency/performance
DeepSeek-V3 is a robust Mixture-of-Experts (MoE) language model developed by DeepSeek, featuring a total of 671 billion parameters, with 37 billion activated per token. It employs Multi-head Latent Attention (MLA) and the DeepSeekMoE architecture to enhance computational efficiency. The model introduces an auxiliary-loss-free load balancing strategy and a multi-token prediction training objective to boost performance. Trained on 14.8 trillion diverse, high-quality tokens, DeepSeek-V3 underwent supervised fine-tuning and reinforcement learning to fully realize its capabilities. Evaluations indicate that it outperforms other open-source models and rivals leading closed-source models, achieving this with a training duration of 55 days on 2,048 Nvidia H800 GPUs, costing approximately $5.58 million. -
4
ML for Trading
Code for machine learning for algorithmic trading, 2nd edition
On over 800 pages, this revised and expanded 2nd edition demonstrates how ML can add value to algorithmic trading through a broad range of applications. Organized in four parts and 24 chapters, it covers the end-to-end workflow from data sourcing and model development to strategy backtesting and evaluation. Covers key aspects of data sourcing, financial feature engineering, and portfolio management. The design and evaluation of long-short strategies based on a broad range of ML algorithms, how to extract tradeable signals from financial text data like SEC filings, earnings call transcripts or financial news. Using deep learning models like CNN and RNN with financial and alternative data, and how to generate synthetic data with Generative Adversarial Networks, as well as training a trading agent using deep reinforcement learning. -
MongoDB Atlas runs apps anywhereMongoDB Atlas gives you the freedom to build and run modern applications anywhere—across AWS, Azure, and Google Cloud. With global availability in over 115 regions, Atlas lets you deploy close to your users, meet compliance needs, and scale with confidence across any geography.
-
5
Weights and Biases
Tool for visualizing and tracking your machine learning experiments
Use W&B to build better models faster. Track and visualize all the pieces of your machine learning pipeline, from datasets to production models. Quickly identify model regressions. Use W&B to visualize results in real time, all in a central dashboard. Focus on the interesting ML. Spend less time manually tracking results in spreadsheets and text files. Capture dataset versions with W&B Artifacts to identify how changing data affects your resulting models. Reproduce any model, with saved code, hyperparameters, launch commands, input data, and resulting model weights. Set wandb.config once at the beginning of your script to save your hyperparameters, input settings (like dataset name or model type), and any other independent variables for your experiments. This is useful for analyzing your experiments and reproducing your work in the future. Setting configs also allows you to visualize the relationships between features of your model architecture or data pipeline and model performance. -
6
MedicalGPT
MedicalGPT: Training Your Own Medical GPT Model with ChatGPT Training
MedicalGPT training medical GPT model with ChatGPT training pipeline, implementation of Pretraining, Supervised Finetuning, Reward Modeling and Reinforcement Learning. MedicalGPT trains large medical models, including secondary pre-training, supervised fine-tuning, reward modeling, and reinforcement learning training. -
7
OpenSpiel
Environments and algorithms for research in general reinforcement
OpenSpiel is a collection of environments and algorithms for research in general reinforcement learning and search/planning in games. OpenSpiel supports n-player (single- and multi- agent) zero-sum, cooperative and general-sum, one-shot and sequential, strictly turn-taking and simultaneous-move, perfect and imperfect information games, as well as traditional multiagent environments such as (partially- and fully- observable) grid worlds and social dilemmas. OpenSpiel also includes tools to analyze learning dynamics and other common evaluation metrics. Games are represented as procedural extensive-form games, with some natural extensions. The core API and games are implemented in C++ and exposed to Python. Algorithms and tools are written both in C++ and Python. To try OpenSpiel in Google Colaboratory, please refer to open_spiel/colabs subdirectory. -
8
Ray
A unified framework for scalable computing
Modern workloads like deep learning and hyperparameter tuning are compute-intensive and require distributed or parallel execution. Ray makes it effortless to parallelize single machine code — go from a single CPU to multi-core, multi-GPU or multi-node with minimal code changes. Accelerate your PyTorch and Tensorflow workload with a more resource-efficient and flexible distributed execution framework powered by Ray. Accelerate your hyperparameter search workloads with Ray Tune. Find the best model and reduce training costs by using the latest optimization algorithms. Deploy your machine learning models at scale with Ray Serve, a Python-first and framework agnostic model serving framework. Scale reinforcement learning (RL) with RLlib, a framework-agnostic RL library that ships with 30+ cutting-edge RL algorithms including A3C, DQN, and PPO. Easily build out scalable, distributed systems in Python with simple and composable primitives in Ray Core. -
9
dm_control
DeepMind's software stack for physics-based simulation
DeepMind's software stack for physics-based simulation and Reinforcement Learning environments, using MuJoCo. DeepMind's software stack for physics-based simulation and Reinforcement Learning environments, using MuJoCo physics. The MuJoCo Python bindings support three different OpenGL rendering backends: EGL (headless, hardware-accelerated), GLFW (windowed, hardware-accelerated), and OSMesa (purely software-based). At least one of these three backends must be available in order render through dm_control. Hardware rendering with a windowing system is supported via GLFW and GLEW. On Linux these can be installed using your distribution's package manager. "Headless" hardware rendering (i.e. without a windowing system such as X11) requires EXT_platform_device support in the EGL driver. While dm_control has been largely updated to use the pybind11-based bindings provided via the mujoco package, at this time it still relies on some legacy components that are automatically generated. -
$300 Free Credits to Build on Google CloudPut your $300 in credit toward real workloads, then keep building with free monthly usage for 20+ products. No commitment and no charge until you upgrade.
-
10
Bullet Physics SDK
Real-time collision detection and multi-physics simulation for VR
This is the official C++ source code repository of the Bullet Physics SDK: real-time collision detection and multi-physics simulation for VR, games, visual effects, robotics, machine learning etc. We are developing a new differentiable simulator for robotics learning, called Tiny Differentiable Simulator, or TDS. The simulator allows for hybrid simulation with neural networks. It allows different automatic differentiation backends, for forward and reverse mode gradients. TDS can be trained using Deep Reinforcement Learning, or using Gradient based optimization (for example LFBGS). In addition, the simulator can be entirely run on CUDA for fast rollouts, in combination with Augmented Random Search. This allows for 1 million simulation steps per second. It is highly recommended to use PyBullet Python bindings for improved support for robotics, reinforcement learning and VR. Use pip install pybullet and checkout the PyBullet Quickstart Guide. -
11
ViZDoom
Doom-based AI research platform for reinforcement learning
ViZDoom allows developing AI bots that play Doom using only the visual information (the screen buffer). It is primarily intended for research in machine visual learning, and deep reinforcement learning, in particular. ViZDoom is based on ZDOOM, the most popular modern source-port of DOOM. This means compatibility with a huge range of tools and resources that can be used to create custom scenarios, availability of detailed documentation of the engine and tools and support of Doom community. Async and sync single-player and multi-player modes. Fast (up to 7000 fps in sync mode, single-threaded). Lightweight (few MBs). Customizable resolution and rendering parameters. Access to the depth buffer (3D vision). Automatic labeling of game objects visible in the frame. Access to the list of actors/objects and map geometry.ViZDoom API is reinforcement learning friendly (suitable also for learning from demonstration, apprenticeship learning or apprenticeship via inverse reinforcement learning. -
12
DouZero
[ICML 2021] DouZero: Mastering DouDizhu
DouZero is a reinforcement learning-based AI for playing DouDizhu, a popular Chinese card game. It focuses on perfecting AI strategies for competitive play using value-based deep RL techniques. -
13
DeepMind Lab
A customizable 3D platform for agent-based AI research
DeepMind Lab is a 3D learning environment based on id Software's Quake III Arena via ioquake3 and other open source software. DeepMind Lab provides a suite of challenging 3D navigation and puzzle-solving tasks for learning agents. Its primary purpose is to act as a testbed for research in artificial intelligence, especially deep reinforcement learning. If you use DeepMind Lab in your research and would like to cite the DeepMind Lab environment, we suggest you cite the DeepMind Lab paper. To enable compiler optimizations, pass the flag --compilation_mode=opt, or -c opt for short, to each bazel build, bazel test and bazel run command. The flag is omitted from the examples here for brevity, but it should be used for real training and evaluation where performance matters. DeepMind Lab ships with an example random agent in python/random_agent.py which can be used as a starting point for implementing a learning agent. -
14
Gymnasium
An API standard for single-agent reinforcement learning environments
Gymnasium is a fork of OpenAI Gym, maintained by the Farama Foundation, that provides a standardized API for reinforcement learning environments. It improves upon Gym with better support, maintenance, and additional features while maintaining backward compatibility. -
15
Machine Learning PyTorch Scikit-Learn
Code Repository for Machine Learning with PyTorch and Scikit-Learn
Initially, this project started as the 4th edition of Python Machine Learning. However, after putting so much passion and hard work into the changes and new topics, we thought it deserved a new title. So, what’s new? There are many contents and additions, including the switch from TensorFlow to PyTorch, new chapters on graph neural networks and transformers, a new section on gradient boosting, and many more that I will detail in a separate blog post. For those who are interested in knowing what this book covers in general, I’d describe it as a comprehensive resource on the fundamental concepts of machine learning and deep learning. The first half of the book introduces readers to machine learning using scikit-learn, the defacto approach for working with tabular datasets. Then, the second half of this book focuses on deep learning, including applications to natural language processing and computer vision. -
16
OSWorld
Benchmarking Multimodal Agents for Open-Ended Tasks
OSWorld is an open-source synthetic world environment designed for embodied AI research and multi-agent learning. It provides a richly simulated 3D world where multiple agents can interact, perform tasks, and learn complex behaviors. OSWorld emphasizes multi-modal interaction, enabling agents to process visual, auditory, and symbolic data for grounded learning in a simulated world. -
17
RL Games
RL implementations
rl_games is a high-performance reinforcement learning framework optimized for GPU-based training, particularly in environments like robotics and continuous control tasks. It supports advanced algorithms and is built with PyTorch. -
18
Stable Baselines3
PyTorch version of Stable Baselines
Stable Baselines3 (SB3) is a set of reliable implementations of reinforcement learning algorithms in PyTorch. It is the next major version of Stable Baselines. You can read a detailed presentation of Stable Baselines3 in the v1.0 blog post or our JMLR paper. These algorithms will make it easier for the research community and industry to replicate, refine, and identify new ideas, and will create good baselines to build projects on top of. We expect these tools will be used as a base around which new ideas can be added, and as a tool for comparing a new approach against existing ones. We also hope that the simplicity of these tools will allow beginners to experiment with a more advanced toolset, without being buried in implementation details. -
19
Unity ML-Agents Toolkit
Unity machine learning agents toolkit
Train and embed intelligent agents by leveraging state-of-the-art deep learning technology. Creating responsive and intelligent virtual players and non-playable game characters is hard. Especially when the game is complex. To create intelligent behaviors, developers have had to resort to writing tons of code or using highly specialized tools. With Unity Machine Learning Agents (ML-Agents), you are no longer “coding” emergent behaviors, but rather teaching intelligent agents to “learn” through a combination of deep reinforcement learning and imitation learning. Using ML-Agents allows developers to create more compelling gameplay and an enhanced game experience. Advancement of artificial intelligence (AI) research depends on figuring out tough problems in existing environments using current benchmarks for training AI models. Using Unity and the ML-Agents toolkit, you can create AI environments that are physically, visually, and cognitively rich. -
20
festival3os
mods to the Festival sokoban solver to run on OSX + Win + linux
Mods to the Festival sokoban solver that allow building on OSX, Linux, & Windows -
21
Atropos
Language Model Reinforcement Learning Environments frameworks
Atropos is a comprehensive open-source framework for reinforcement learning (RL) environments tailored specifically to work with large language models (LLMs). Designed as a scalable ecosystem of environment microservices, Atropos allows researchers and developers to collect, evaluate, and manage trajectories (sequences of actions and outcomes) generated by LLMs across a variety of tasks—from static dataset benchmarks to dynamic interactive games and real-world scenario environments. It provides foundational tooling for asynchronous RL loops where environment services communicate with trainers and inference engines, enabling complex workflow orchestration in distributed and parallel setups. This framework facilitates experimentation with RLHF (Reinforcement Learning from Human Feedback), RLAIF, or multi-turn training approaches by abstracting environment logic, scoring, and logging into reusable components. -
22
TorchRL
A modular, primitive-first, python-first PyTorch library
TorchRL is an open-source Reinforcement Learning (RL) library for PyTorch. TorchRL provides PyTorch and python-first, low and high-level abstractions for RL that are intended to be efficient, modular, documented, and properly tested. The code is aimed at supporting research in RL. Most of it is written in Python in a highly modular way, such that researchers can easily swap components, transform them, or write new ones with little effort. -
23
robosuite
A Modular Simulation Framework and Benchmark for Robot Learning
Robosuite is a modular and extensible simulation framework for robotic manipulation tasks, built on top of MuJoCo. Developed by the ARISE Initiative, Robosuite offers a set of standardized benchmarks and customizable environments designed to advance research in robotic manipulation, control, and imitation learning. It emphasizes realistic simulations and ease of use for both single-task and multi-task learning. -
24
verl
Volcano Engine Reinforcement Learning for LLMs
VERL is a reinforcement-learning–oriented toolkit designed to train and align modern AI systems, from language models to decision-making agents. It brings together supervised fine-tuning, preference modeling, and online RL into one coherent training stack so teams can move from raw data to aligned policies with minimal glue code. The library focuses on scalability and efficiency, offering distributed training loops, mixed precision, and replay/buffering utilities that keep accelerators busy. It ships with reference implementations of popular alignment algorithms and clear examples that make it straightforward to reproduce baselines before customizing. Data pipelines treat human feedback, simulated environments, and synthetic preferences as interchangeable sources, which helps with rapid experimentation. VERL is meant for both research and production hardening: logging, checkpointing, and evaluation suites are built in so you can track learning dynamics and regressions over time. -
25
Highly modularized Reinforcement Learning library for real/simulation robots to learn behaviors. Our ultimate goal is to develop an artificial intelligence (AI) program with which the robots can learn to behave as their users wish.
Open Source Reinforcement Learning Frameworks Guide
Open source reinforcement learning frameworks provide developers, researchers, and organizations with the tools needed to create, train, evaluate, and deploy reinforcement learning models. These frameworks support environments where an intelligent agent learns by interacting with its surroundings and improving its decision-making through rewards and penalties. By making the underlying source code publicly available, they enable greater transparency, customization, and collaboration while allowing users to adapt the framework to a wide range of machine learning projects.
Many open source reinforcement learning frameworks include features for environment simulation, policy optimization, model evaluation, distributed training, and experiment management. They often support multiple reinforcement learning algorithms, making it easier to compare different approaches and refine model performance over time. Integration with machine learning libraries, cloud infrastructure, visualization tools, and data processing platforms also helps streamline development workflows for both research and production environments.
As reinforcement learning continues to expand into industries such as robotics, manufacturing, finance, healthcare, logistics, and autonomous systems, open source frameworks have become an important part of AI development. Organizations can use them to accelerate experimentation, reduce development costs, and build solutions tailored to specific operational goals. Their flexibility and active community contributions make them valuable resources for teams seeking to advance reinforcement learning initiatives while maintaining control over their development process.
What Features Do Open Source Reinforcement Learning Frameworks Provide?
- Flexible training pipelines: Build, test, and refine reinforcement learning workflows for different environments and learning objectives.
- Multiple algorithm support: Access diverse reinforcement learning methods to address varied control, optimization, and decision-making challenges.
- Environment compatibility: Connect with simulation environments for consistent training, testing, and performance evaluation.
- Custom model configuration: Adjust architectures, hyperparameters, and learning settings to match unique project requirements.
- Distributed training capabilities: Accelerate large experiments by spreading workloads across multiple processors or computing resources.
- Performance monitoring: Track rewards, learning progress, and evaluation metrics throughout model development.
- Checkpoint management: Save and restore training states to continue experiments without restarting from the beginning.
- Hardware acceleration: Leverage supported processing hardware to improve training efficiency and reduce execution time.
- Experiment reproducibility: Maintain consistent configurations, random seeds, and settings for repeatable testing and reliable comparisons.
Different Types of Open Source Reinforcement Learning Frameworks
- General-purpose frameworks: Support multiple reinforcement learning algorithms across diverse environments and research objectives.
- Deep reinforcement learning frameworks: Focus on neural network-based learning for solving complex decision-making tasks.
- Distributed training frameworks: Enable scalable model training across multiple devices or computing resources.
- Multi-agent reinforcement learning frameworks: Help train multiple agents that cooperate, compete, or interact within shared environments.
- Offline reinforcement learning frameworks: Learn policies from previously collected datasets without requiring continuous environment interaction.
- Model-based reinforcement learning frameworks: Build predictive environment models to improve planning efficiency and reduce training requirements.
- Robotics-focused frameworks: Emphasize control, simulation, and physical interaction for autonomous robotic applications.
What Are the Advantages Provided by Open Source Reinforcement Learning Frameworks?
- Flexible customization: Developers can modify architectures, algorithms, and workflows to match specialized research or production requirements.
- Cost efficiency: Organizations reduce licensing expenses while expanding experimentation across multiple projects.
- Community collaboration: Shared improvements accelerate innovation and encourage continuous enhancements from contributors worldwide.
- Broad algorithm support: Multiple reinforcement learning approaches simplify testing different strategies for varied objectives.
- Transparency: Accessible source code helps teams understand model behavior, implementation details, and optimization opportunities.
- Integration capabilities: Frameworks connect with machine learning, simulation, analytics, and infrastructure tools for streamlined development.
- Scalability: Teams can expand training workloads across different computing environments as project demands increase.
- Faster experimentation: Built-in components reduce development time while simplifying iterative testing and performance evaluation.
Who Uses Open Source Reinforcement Learning Frameworks?
- AI researchers: Build, evaluate, and refine reinforcement learning methods using customizable open source frameworks.
- Academic institutions: Support teaching, experimentation, and collaborative research across reinforcement learning topics.
- Machine learning engineers: Develop intelligent decision-making systems for production and testing environments.
- Robotics teams: Train autonomous machines to improve movement, navigation, and task completion.
- Data science teams: Explore sequential decision models to solve complex optimization challenges.
- Research laboratories: Conduct controlled experiments while comparing reinforcement learning approaches and configurations.
- Product development teams: Create adaptive features that respond intelligently to changing user interactions.
- Autonomous vehicle developers: Simulate driving decisions before validating models in real-world environments.
How Much Do Open Source Reinforcement Learning Frameworks Cost?
Open source reinforcement learning frameworks are available across a wide range of pricing models, with many offering their core capabilities at no licensing cost. Organizations can download, modify, and deploy these frameworks without paying subscription fees, making them an attractive option for research institutions, startups, and development teams. However, the overall cost depends on factors such as infrastructure, computing resources, implementation complexity, and ongoing maintenance. Larger projects that require distributed training, extensive experimentation, or high-performance hardware may incur significantly higher operational expenses.
Although the framework itself may be free to use, businesses should account for indirect costs when planning adoption. Expenses may include cloud computing, specialized hardware, technical support, staff training, integration with existing tools, and model optimization. Some organizations also invest in consulting services or managed infrastructure to accelerate deployment and improve reliability. Evaluating both the upfront and long-term operational costs provides a more accurate understanding of the total investment required.
What Do Open Source Reinforcement Learning Frameworks Integrate With?
Open source reinforcement learning frameworks can integrate with machine learning platforms that manage model training, experimentation, and deployment. They also work with data processing tools that prepare training datasets, transform input data, and support feature engineering. Many frameworks connect with simulation environments, robotics platforms, and game engines that provide interactive environments for training and evaluating reinforcement learning agents. Integration with cloud infrastructure and container orchestration platforms allows teams to scale training workloads, manage resources, and automate deployment pipelines.
These frameworks can also connect with monitoring and observability tools that track training performance, resource usage, and experiment results. Version control platforms, workflow automation tools, and continuous integration and delivery solutions help streamline collaboration and model lifecycle management. Database systems, analytics platforms, and visualization tools support data storage, reporting, and performance analysis. Security, identity management, and API management solutions further improve governance by controlling access, securing integrations, and simplifying communication between connected applications.
What Are the Trends Relating to Open Source Reinforcement Learning Frameworks?
- Better support for large-scale distributed training improves experimentation speed and model scalability.
- Expanded compatibility with modern AI ecosystems simplifies development across multiple machine learning workflows.
- Improved benchmarking methods encourage more reliable performance comparisons across different reinforcement learning approaches.
- Growing interest in robotics increases demand for realistic simulation and training environments.
- More frameworks emphasize modular architectures that simplify customization and maintenance.
- Enhanced support for offline reinforcement learning broadens practical business applications.
- Increased focus on reproducibility encourages standardized evaluation methods and consistent research outcomes.
Getting Started With Open Source Reinforcement Learning Frameworks
Selecting the right open source reinforcement learning framework starts with identifying your project's objectives, deployment environment, and team expertise. Consider whether the framework supports the algorithms, environments, and training methods required for your use case. Evaluate compatibility with your existing technology stack, hardware, and machine learning workflows to minimize integration challenges.
Performance and scalability are equally important. Look for frameworks that efficiently handle distributed training, large datasets, and hardware acceleration if your workloads demand significant computing resources. Documentation quality, update frequency, community activity, and long-term maintenance can also influence productivity and future reliability. Finally, review licensing terms, customization options, debugging capabilities, and available evaluation tools to ensure the framework aligns with your organization's technical, operational, and compliance requirements. Testing several frameworks with a small proof of concept can provide valuable insight before making a long-term decision.
