Fast, accurate physics engine for contact-rich robotics: soft-contact convex solver trusted for manipulation and locomotion research, with MJX (JAX) for massively parallel GPU/TPU simulation — Apache-2.0 since DeepMind open-sourced it.
Fast, accurate physics engine for contact-rich robotics: soft-contact convex solver trusted for manipulation and locomotion research, with MJX (JAX) for massively parallel GPU/TPU simulation — Apache-2.0 since DeepMind open-sourced it.
Save it, compare it, request a quote or record that you have used it — with an account.
MuJoCo (Multi-Joint dynamics with Contact) is a physics engine purpose-built for fast, accurate simulation of articulated systems in contact — the property that made it the default for manipulation, locomotion and reinforcement-learning research. Its convex soft-contact model delivers stable, differentiable-friendly dynamics; models are described in MJCF XML; and MJX reimplements the pipeline in JAX for thousands of parallel environments on GPU/TPU. Acquired and open-sourced by DeepMind (Apache-2.0) in 2022, it is under active development (3.x line) with Python bindings, interactive viewer, and the Menagerie library of validated robot models.
This product is part of the Birdwave Atlas community.
Contribute evaluations, share integrations and help improve compatibility data.
Discussions, evaluations, integrations and updates