Skip to main content

April 27, 2016

OpenAI Gym Beta

Openai Gym Beta

We’re releasing the public beta of OpenAI Gym, a toolkit for developing and comparing reinforcement learning (RL) algorithms. It consists of a growing suite of environments (from simulated robots to Atari games), and a site for comparing and reproducing results.

OpenAI Gym is compatible with algorithms written in any framework, such as Tensorflow(opens in a new window) and Theano(opens in a new window). The environments are written in Python, but we’ll soon make them easy to use from any language. We originally built OpenAI Gym as a tool to accelerate our own RL research. We hope it will be just as useful for the broader community.

Getting started

If you’d like to dive in right away, you can work through our tutorial(opens in a new window). You can also help out while learning by reproducing a result(opens in a new window).

Why RL?

Reinforcement learning (RL) is the subfield of machine learning concerned with decision making and motor control. It studies how an agent can learn how to achieve goals in a complex, uncertain environment. It’s exciting for two reasons:

However, RL research is also slowed down by two factors:

  • The need for better benchmarks. In supervised learning, progress has been driven by large labeled datasets like ImageNet(opens in a new window). In RL, the closest equivalent would be a large and diverse collection of environments. However, the existing open-source collections of RL environments don’t have enough variety, and they are often difficult to even set up and use.

  • Lack of standardization of environments used in publications. Subtle differences in the problem definition, such as the reward function or the set of actions, can drastically alter a task’s difficulty. This issue makes it difficult to reproduce published research and compare results from different papers.

OpenAI Gym is an attempt to fix both problems.

The environments

OpenAI Gym provides a diverse suite of environments that range from easy to difficult and involve many different kinds of data. We’re starting out with the following collections:

Over time, we plan to greatly expand this collection of environments. Contributions from the community are more than welcome.

Each environment has a version number (such as Hopper-v0(opens in a new window)). If we need to change an environment, we’ll bump the version number, defining an entirely new task. This ensures that results on a particular environment are always comparable.


We’ve made it easy to upload results(opens in a new window) to OpenAI Gym. However, we’ve opted not to create traditional leaderboards. What matters for research isn’t your score (it’s possible to overfit or hand-craft solutions to particular tasks), but instead the generality of your technique.

We’re starting out by maintaining a curated list(opens in a new window) of contributions that say something interesting about algorithmic capabilities. Long-term, we want this curation to be a community effort rather than something owned by us. We’ll necessarily have to figure out the details over time, and we’d would love your help(opens in a new window) in doing so.

We want OpenAI Gym to be a community effort from the beginning. We’ve starting working with partners to put together resources around OpenAI Gym:

During the public beta, we’re looking for feedback on how to make this into an even better tool for research. If you’d like to help, you can try your hand at improving the state-of-the-art on each environment, reproducing other people’s results, or even implementing your own environments. Also please join us in the community chat(opens in a new window)!