MAGMA-BENCH

Evaluate agents on interactive robotic tasks with MAGMA. MAGMA-BENCH loads compiled benchmark episodes, runs them in simulation, communicates with an agent through the MAGMA HTTP protocol, and saves detailed results and metrics.

Version 2.0.0b3 is a beta of v2. APIs, command-line options, and result formats may change before the stable release. End-to-end validation is still in progress.

Official documentation · Report an issue

Installation

Python 3.12 is required. Simulation installation has been checked on Linux x86_64. GPU simulation and rendering require compatible system drivers.

python -m pip install "magma_bench==2.0.0b3"

This installs magma_core[simulation]>=2.0.0b1,<3.0.0 and magma_scenarios>=2.0.1,<3.0.0 with their Python dependencies. The benchmark files and the agent server are separate inputs and are not bundled with this package.

The version is pinned explicitly because this is a prerelease. To request the latest version, including prereleases, use:

python -m pip install --upgrade --pre magma_bench

Usage

Start a compatible MAGMA agent server, then run a compiled benchmark directory:

magma-bench run \
  --benchmark-root /path/to/compiled-benchmark \
  --agent-address http://127.0.0.1:8888 \
  --run-name experiment-1

The agent must expose /health, /v1/info, and /v1/responses using protocol version 2.0. Configuration can be supplied with --config-path; command-line options override its benchmark settings.

Useful options include:

magma-bench run --help
magma-bench run --benchmark-root /path/to/benchmark --scenarios coffee_comp
magma-bench run --benchmark-root /path/to/benchmark --skip-judge
magma-bench run --benchmark-root /path/to/benchmark --model-logs --videos all
magma-bench run --benchmark-root /path/to/benchmark --videos planner-failure

Video output uses the optional dependencies:

python -m pip install "magma_bench[video]==2.0.0b3"

Version 2 uses result schema 2.0. Resume a compatible run with --results-path; older result schemas and agent adapters are unsupported.

See the official documentation for benchmark preparation, configuration, agent setup, metrics, and result formats.

Install from source

Install the tagged release from GitHub:

python -m pip install "magma_bench @ git+https://github.com/MAGMA-rob/magma-bench.git@v2.0.0b3"

Dependencies are resolved from PyPI. For local development, clone the repository and run python -m pip install -e ".[dev,video]".

License

BSD 2-Clause.