MoLFormer-XL

Open weights IBM January 2023

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
IBM
Organisation type
Industry
Country
United States of America
Published
25 January 2023
Authors
Jerret Ross, Brian Belgodere, Vijil Chenthamarakshan, Inkit Padhi, Youssef Mroueh, Payel Das

What it does

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

Domain
Biology
Task
Protein generation, Protein folding prediction

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

"MoLFormer-XL has been pretrained on 1.1 billion molecules represented as machine-readable strings of text." 15K steps "PubChem+ZINC (>1 billion data points) datasets" mean token amount (table 4 from supplementary materials): 44.76 44.76*1.1*10^9 = 49236000000 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
4.5 × 10²⁰ FLOP

125000000000000 FLOP / GPU / sec *(208 hours * 16 GPUs +1 GPU * 12 hours) [see training time notes] * 3600 sec / hour * 0.3 [assumed utilization] = 4.509e+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
Chips used
16
Wall-clock time
208 hours (8.7 days)

"Together, both techniques raised our per-GPU processing costs from 50 molecules to 1,600 molecules, allowing us to get away with 16 GPUs instead of 1,000. By eliminating hundreds of unnecessary GPUs, we consumed 61 times less energy and still had a trained model in five days. " "Our pretraining task consists of training on the full dataset to 4 epochs. Training a single epoch of just PubChem on a single NVIDIA V100 GPU would take approximately 60 hours. Utilizing Distributed Data Parallel, pre…

Power draw
9.6 kW

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

Python codes for MoLFormer training and fine-tuning, and Python notebooks for MoLFormer attention visualization, as well as instances of pretrained models, are available at https://github.com/IBM/molformer. Apache 2.0

How it is classified

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

Record confidence
Confident

Sources

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

Reference
An AI foundation model that learns the grammar of molecules
Last updated
28 November 2025

What the numbers mean

About this model

MoLFormer-XL was published by IBM, in United States of America, in January 2023. It comes out of industry.

It works in Biology, and is recorded as doing protein generation, Protein folding prediction.

The weights being open is what puts this page in the calculator rather than only in the catalogue: it is a model you can actually hold.

Training and provenance

The training run consumed about 4.5 × 10²⁰ FLOP, on NVIDIA V100. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Answers

MoLFormer-XL — common questions

01

What is MoLFormer-XL used for?

MoLFormer-XL works in Biology, and is recorded as handling protein generation, Protein folding prediction. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

02

Where can I download MoLFormer-XL?

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

03

How much compute was used to train MoLFormer-XL?

Around 4.5 × 10²⁰ FLOP, on 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.

04

What GPU do I need to run MoLFormer-XL?

We cannot say. MoLFormer-XL 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.

05

Is MoLFormer-XL open source?

Its weights are published, so MoLFormer-XL 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.

06

How many parameters does MoLFormer-XL have?

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

07

Who created MoLFormer-XL?

MoLFormer-XL was published by IBM, based in United States of America, categorised as industry.

08

When was MoLFormer-XL released?

MoLFormer-XL was published in January 2023. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

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.