GemNet-OC

Open weights Technical University of Munich,Carnegie Mellon University (CMU),Facebook AI Research September 2022

No estimate

No hardware requirements for this model

This model's weights are open, but no parameter count has been published for it. Every memory and speed figure starts from that number, so we would rather show nothing than a fabricated estimate.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation
Technical University of Munich,Carnegie Mellon University (CMU),Facebook AI Research
Organisation type
Academia,Academia,Industry
Country
Germany, United States of America, France
Published
30 September 2022
Authors
Johannes Gasteiger, Muhammed Shuaibi, Anuroop Sriram, Stephan Günnemann, Zachary Ulissi, C. Lawrence Zitnick, Abhishek Das

What it does

The problem areas the model was built for. A model can carry several of each.

Domain
Materials science
Task
Molecular simulation, Molecular property prediction, Atomistic simulations, Molecular representation learning, Materials design

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Training data
tokens

"trained for a max number of epochs (4) or until the learning rate has been exhaustively stepped, whichever comes first" from the Table 1: ∼134 M samples (133 934 018) avg size 73.3 (7-225) 133 934 018 * 73.3 = 9817363519.4 atoms (~tokens)

Epochs
4

Training compute

The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.

Training compute
5.4 × 10²⁰ FLOP

125000000000000 FLOP / GPU / sec [bf16 assumed] * 4000 GPU-hours [inferring from Figure 2] * 3600 sec / hour * 0.3 [asssumed utilization] = 5.4e+20 FLOP

How it was established
Hardware

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Training hardware
NVIDIA V100
Chip-hours
4,000

Availability

Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.

Weights
Open — downloadable
Model access
Open weights (unrestricted)
Training code
Open source

MIT license https://github.com/facebookresearch/fairchem/tree/main/src/fairchem/core

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Record confidence
Likely

Sources

Where this record came from and when it was last checked.

Reference
GemNet-OC: Developing Graph Neural Networks for Large and Diverse Molecular Simulation Datasets
Last updated
28 November 2025

What the numbers mean

About this model

GemNet-OC was published by Technical University of Munich,Carnegie Mellon University (CMU),Facebook AI Research, in the country recorded as Germany, during September 2022. It comes out of an organisation categorised as academia,Academia,Industry.

It works in the domain of Materials science, and is recorded as performing the task of molecular simulation, Molecular property prediction, Atomistic simulations, Molecular representation learning, Materials design.

Because its weights were released, nothing about running it depends on a provider staying available — it is yours once downloaded.

How it was trained

Producing it required arithmetic totalling around 5.4 × 10²⁰ FLOP, on hardware recorded as NVIDIA V100. That figure measures what producing the model cost, and has no bearing on how fast it answers.

Answers

GemNet-OC — common questions

01

GemNet-OC— how much compute was used to train it?

Training consumed around 5.4 × 10²⁰ FLOP, on hardware recorded as NVIDIA V100. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.

02

GemNet-OC— what GPU do I need to run it?

We cannot say. It has open weights, but no parameter count has been published for it, and every memory and speed calculation starts from that number. We would rather show nothing than a fabricated estimate.

03

GemNet-OC— is it open source?

Its weights are published, so it can be downloaded and run on your own hardware. Note that open weights is not the same as open source in the full sense — it says nothing about the training data, the training code, or the commercial terms attached.

04

GemNet-OC— how many parameters does it have?

No parameter count has been published for it, which is why no memory or speed figure appears on this page.

05

GemNet-OC— who created it?

It was published by Technical University of Munich,Carnegie Mellon University (CMU),Facebook AI Research, based in Germany, an organisation categorised as academia,Academia,Industry.

06

GemNet-OC— when was it released?

It was published in September 2022. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

07

GemNet-OC— what is it used for?

It works in the domain of Materials science, and is recorded as handling the task of molecular simulation, Molecular property prediction, Atomistic simulations, Molecular representation learning, Materials design. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

08

GemNet-OC— where can I download it?

The weights are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

Source

Original publication

Record last updated 28 November 2025

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.