# newton-physics/newton

An open-source, GPU-accelerated physics simulation engine built upon NVIDIA Warp, specifically targeting roboticists and simulation researchers.

Repository: https://github.com/newton-physics/newton
Canonical: https://ross.abutalabs.com/products/newton
Homepage: https://newton-physics.github.io/newton/stable/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: physics-simulation, robotics, nvidia-warp, newton-physics
Last push: 2026-08-26T23:12:25+00:00

## Health v2 (maintenance only)
Score: 86/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 97, longevity 35
- inputs: {"age_days": 498, "days_push": 7, "days_rel": 22, "gap_med": 24.0, "n_releases_24m": 7}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5539, forks 658 (observed 2026-08-28T04:09:22.095243+00:00)

## What it is
Newton is an open-source, GPU-accelerated physics simulation engine built on NVIDIA Warp, targeting roboticists and simulation researchers. It extends Warp's deprecated sim module, integrates MuJoCo Warp as its primary backend, and emphasizes GPU computation, OpenUSD support, and differentiability.

## Use cases
- simulate robots on GPU at scale
- run differentiable physics simulation for reinforcement learning
- train robot policies in simulation
- simulate articulated bodies and contact dynamics
- build USD-based robotics simulation pipelines
- benchmark physics solvers like MuJoCo Warp

## When to choose
- you need GPU-parallel physics simulation for robotics research
- you want differentiable simulation integrated with Python
- you need OpenUSD scene support in a physics engine
- you want a community-governed, Apache-2.0 physics engine backed by Disney Research, Google DeepMind, and NVIDIA

## When to avoid
- you need CPU-only simulation on macOS with GPU acceleration
- you need a mature, long-stable engine with a large ecosystem like MuJoCo or Isaac Sim standalone
- your project does not involve robotics or physics simulation
- you lack an NVIDIA GPU (Maxwell or newer) for full performance

## Facets
- artifact type: library
- maturity: active
- function: simulation, machine-learning, robotics, gpu-computing
- domain: robotics, simulation, gpu-computing, machine-learning
- platform: windows, python
- tags: physics-engine, nvidia-warp, mujoco, openusd, differentiable-simulation, robotics-simulation, linux, macos, gpu

## Member repositories
- newton-physics/newton (main) score 86

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:22.095243+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 2026-08-29T17:55:42.928644+00:00, confidence not recorded.
  - readme: https://github.com/newton-physics/newton (fetched 2026-08-28T04:09:22.095243+00:00, sha b36c2731a009)
  - homepage: https://newton-physics.github.io/newton/stable/ (fetched 2026-08-29T08:51:02.886357+00:00, sha e402629ca176)
- Data as of 2026-08-30T08:39:29.467469+00:00.