Project AURIONAI Use-Cases Research & Innovation
materials

PET-OAM-XL

Machine-learned interatomic potential on Matbench Discovery (rank 1 by CPS, date added 2026-01-10).

EPFLEPFL
EPFL
Levelmodel
Classmaterials
Parameters730M
Contextnot in the notes
Released10 Jan 2026
Opennessunknown
Runs onnot in the notes
Statusactive

Benchmark Results

maintainer-run
Matbench Discoverydomain
maintainer-run
0.924
00.51

Score in F1. Higher is better.

leaderboard fetched 2026-10-1010 Jan 2026 matbench-discovery.materialsproject.org
maintainer-run
0.898
00.450.9

Score in CPS. Higher is better.

rank 1, leaderboard fetched 2026-10-10, ranked by CPS10 Jan 2026 matbench-discovery.materialsproject.org
lower is bettermaintainer-run
0.119
00.1250.25

Score in kSRME. Lower is better.

leaderboard fetched 2026-10-1010 Jan 2026 matbench-discovery.materialsproject.org
Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).

R&D Topics

Universal machine-learned interatomic potentialsexemplar

Identity

Developer
EPFL
Family
not in the notes
Released
10 Jan 2026 sourcesregistry date
Weights published
–
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
GNN Equivariant GNN interatomic potential (MLIP)
Total parameters
730M sources
Active parameters
–
Checkpoint parameters
–
Note
Training data OMat24+sAlex+MPtrj
Context length
not in the notes
Input
molecule
Output
molecule
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
unknown
Licence
not in the notes sources
Access routes
not in the notes
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
not in the notes
Footprint
not in the notes
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
not stated
Regional languages
not in the notes
Jakarta hosting
unknown
Languages
not in the notes
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Lineage

The notes record no lineage for this model.

Industry Use Cases

No industry use case links this model.

Sources

  1. Matbench Discovery matbench-discovery.materialsproject.org https://matbench-discovery.materialsproject.org/
  2. A new reference model for machine learning-driven materials discovery actu.epfl.ch https://actu.epfl.ch/news/a-new-reference-model-for-machine-learning-driven

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "PET-OAM-XL", https://aurion.semai.id/models/epfl-pet-oam-xl/, records of October 2026.