microsoft/mattergen

Official implementation of MatterGen -- a generative model for inorganic materials design across the periodic table that can be fine-tuned to steer the generation towards a wide range of property constraints.

74
/ 100
Verified

Built on diffusion models with classifier-free guidance, MatterGen generates crystal structures by iteratively denoising atomic coordinates and types conditioned on periodic table composition. The framework includes multiple pre-trained checkpoints for unconditional and property-conditioned generation (magnetic density, band gap, bulk modulus, etc.), with outputs in CIF and extended XYZ formats, and integrates MatterSim for structure relaxation and evaluation metrics like novelty and stability.

1,643 stars and 184 monthly downloads. Available on PyPI.

Maintenance 10 / 25
Adoption 15 / 25
Maturity 25 / 25
Community 24 / 25

How are scores calculated?

Stars

1,643

Forks

308

Language

Python

License

MIT

Last pushed

Feb 27, 2026

Monthly downloads

184

Commits (30d)

0

Dependencies

35

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/generative-ai/microsoft/mattergen"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.