DeepSWE benchmarks coding agents on original, long-horizon software engineering tasks.
Open source
24 alternativesmetaworld
MetaWorld provides robotics environments for benchmarking multi task and meta reinforcement learning.
About metaworld
Meta-World is an open source benchmark for developing and evaluating multi-task and meta reinforcement learning algorithms. Its environments are continuous control robotic manipulation tasks. It includes multi-task benchmarks (MT1, MT10, MT50) for learning a policy across one, 10, or 50 training tasks, and meta-learning benchmarks (ML1, ML10, ML45) for testing few-shot adaptation. The API follows the Gymnasium interface for creating and interacting with environments.
The project is installed with pip and is tested on Python 3.11, 3.12, and 3.13 on Linux and macOS. Windows pull requests are accepted, but Windows is not officially supported. MT10 and MT50 offer synchronous and asynchronous vector environment options. The synchronous option needs the least compute, while the asynchronous option is intended for users with more compute. Documentation is hosted at metaworld.farama.org.
Key features
- Multi-task benchmarks MT1, MT10, and MT50 for training multi-task policies
- Meta-learning benchmarks ML1, ML10, and ML45 for few-shot adaptation testing
- Gymnasium-compatible API for environment creation and interaction
- Synchronous and asynchronous vector environment options for MT10 and MT50
- One-hot task IDs appended to observations in MT10 and MT50 benchmarks
- Seed parameter for reproducible environment creation
Best for: Reinforcement learning researchers who need a standard benchmark for evaluating multi-task and meta reinforcement learning algorithms on robotic manipulation tasks.
Good for
- Open source benchmark with several settings for evaluating different aspects of reinforcement learning algorithms
- Follows the Gymnasium API for environment creation and interaction
- Synchronous vector option needs the least compute resources
Watch out for
- Windows is not officially supported; pull requests for it are accepted
- Officially tested only on Python 3.11, 3.12, and 3.13 on Linux and macOS
metaworld at a glance
- Category
- AI & Machine Learning › Models & ML Research
- Type
- Open-source project
- Pricing
- Not confirmed. Check the site
- Open source
- Yes · Source code
- Language
- English
- Last checked
- Oct 6, 2026 · Online when checked
- Listed
- Oct 5, 2026
metaworld alternatives
Tools and sites you could use instead, ranked by how closely they match.
The learn2learn library supports meta learning experiments and research with PyTorch.
Open source
24 alternatives
OpenVLA provides an open-source vision-language-action model for robotic manipulation tasks.
Open source
24 alternatives
DeepSeek Coder is a language model project for code generation and programming tasks.
Open source
24 alternatives
The BELLE project provides an open-source Chinese-language conversational model.
Open source
24 alternatives
SWE-bench evaluates whether language models can resolve real GitHub issues.
Open source
24 alternativesPopular in Models & ML Research
Well-known listings from the same topic.
Falcon LLMfalconllm.tii.aeERNIEernie.baidu.comTokenVersetoken-verse.github.iowTAObittensor.comAlpacacrfm.stanford.edu/2023/03/13/alpaca.htmlSmolVLAhuggingface.co/blog/smolvla
metaworld FAQ
How is Meta-World installed?
It is installed with the command pip install metaworld.
Which Python versions and operating systems are tested?
The project tests Python 3.11, 3.12, and 3.13 on Linux and macOS. Windows is not officially supported.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.