# google-deepmind/mujoco_mpc

Real-time behaviour synthesis with MuJoCo, using Predictive Control

Repository: https://github.com/google-deepmind/mujoco_mpc
Canonical: https://ross.abutalabs.com/products/mujoco_mpc
Homepage: https://github.com/deepmind/mujoco_mpc
Language: C++
License: Apache-2.0
License Family: permissive
Topics: model-predictive-control, mpc, mpc-control, mujoco, predictive-control, motor-control
Last push: 2026-08-21T13:49:06+00:00

## Health v2 (maintenance only)
Score: 66/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 98, release rhythm 8, longevity 97
- inputs: {"age_days": 1371, "days_push": 12, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1713, forks 281 (observed 2026-08-28T04:05:26.031845+00:00)

## What it is
MuJoCo MPC (MJPC) is an interactive application and C++ software framework from Google DeepMind for real-time predictive control with the MuJoCo physics simulator. It supports multiple shooting-based planners including derivative-based methods (iLQG, Gradient Descent) and a derivative-free Predictive Sampling planner, with a GUI for authoring and solving robotics tasks.

## Use cases
- solve quadruped locomotion tasks with model-predictive control
- control bimanual robotic manipulation in simulation
- track humanoid motion capture in real time
- author and solve complex robotics tasks in MuJoCo
- compare shooting-based MPC planners like iLQG and Predictive Sampling
- experiment with real-time behaviour synthesis for robots

## When to choose
- you need real-time model-predictive control with MuJoCo
- you want a GUI to interactively author and solve robotics control tasks
- you need shooting-based planners such as iLQG, Gradient Descent, or Predictive Sampling
- you are doing robotics research on locomotion, manipulation, or motion tracking

## When to avoid
- you need a lightweight headless control library without a GUI
- your project does not use MuJoCo as its physics engine
- you need sampling-based motion planning rather than predictive control
- you require Windows support, which is not a tested platform

## Facets
- artifact type: application
- maturity: active
- function: simulation, robotics, gui, developer-tools
- domain: robotics, simulation, machine-learning
- platform: cpp, python
- tags: model-predictive-control, mujoco, ilqg, predictive-sampling, robotics-tasks, real-time-control, linux, macos

## Member repositories
- google-deepmind/mujoco_mpc (main) score 66

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:26.031845+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:34:42.692742+00:00, confidence not recorded.
  - readme: https://github.com/google-deepmind/mujoco_mpc (fetched 2026-08-28T04:05:26.031845+00:00, sha 21ea9be33bb1)
  - homepage: https://github.com/deepmind/mujoco_mpc (fetched 2026-08-29T11:10:39.620829+00:00, sha 31a23f2367b0)
- Data as of 2026-08-30T08:39:29.467469+00:00.