Chemistry & ADMET · mace-foundations
MACE-MPA-0
MACE foundation model predicting energy, forces, and charges for inorganic materials and molecules.
What MACE-MPA-0 is
MACE-MPA-0 is a machine learning foundation model developed by mace-foundations for inorganic-materials simulation and computational chemistry tasks.
The model offers a broader coverage variant for simulating inorganic materials and predicting atomic-scale properties. Taking molecular representations as inputs, it calculates core mechanical and electronic properties including total energy, maximum atomic forces, and partial charges across individual atoms.
How it works
MACE-MPA-0 uses the MACE architecture and was trained on combined reference data from the Materials Project and the Alexandria datasets to expand structural and compositional coverage for inorganic materials.
Key facts
- Domain
- Chemistry, materials science, physics
- Training data
- Materials Project and Alexandria datasets
- Input format
- SMILES strings
- Output format
- Tabular properties (energy, maximum force, per-atom partial charges)
- GPU memory requirement
- 16 GB
What you can use it for
- Inorganic Materials Simulation
- Simulating inorganic crystalline and solid-state systems using parameters trained on comprehensive materials databases.
- Molecular Energy Calculation
- Predicting the potential energy of chemical structures supplied as SMILES representations.
- Atomic Force Estimation
- Evaluating maximum forces acting on atomic configurations for structural assessment.
- Partial Charge Assignment
- Calculating per-atom partial charges to analyse electrostatic distributions within molecular systems.
Run it here
No installation and no GPU of your own: fill in the inputs and it runs on our servers. Results stay in your browser unless you sign in.
Runs on On-demand GPU
Runs on our servers with on-demand compute. A first run needs time to load the model; active capacity can be reused and scales down when idle.Run this model
MACE-MPA-0
Strengths and limits
Good at
- Provides broader coverage for inorganic systems by combining the Materials Project and Alexandria datasets during training.
- Predicts energy, maximum force, and per-atom partial charges simultaneously from structural input.
Watch out for
- Requires substantial hardware resources, needing at least 16 GB of GPU memory.
- Interface specifications for inputs and outputs are derived from generic pipeline tags and require validation before relying on computational outputs.
Inputs and outputs
You provide
Molecule (SMILES)
We generate and relax a 3D conformer before the model sees it.
You get back
Energy, maximum force and per-atom partial charges.
Licence
Check the licence before commercial use · See model card
The See model card licence carries terms worth reading before you depend on it.
Questions
Can I use MACE-MPA-0 commercially?
Check the licence before commercial use. The See model card licence carries terms worth reading before you depend on it.
How do I run MACE-MPA-0 without installing anything?
Use the form under “Run it here” on this page: fill in the inputs and press Run model. It runs on our servers on CPU only, typically in 12s, and a run costs about 28 credits. Visitors without an account get 1,000 credits a month.
What datasets was MACE-MPA-0 trained on?
It was trained on data from both the Materials Project and the Alexandria datasets.
What inputs does the model require?
The model accepts molecular structures represented as SMILES strings.
Which physical properties does the model output?
It outputs a table containing energy, maximum force, and per-atom partial charges.
Written from MACE-MPA-0's own documentation (model card). Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.