Multi-Modal Foundation · arcinstitute
SE-600M
Generates embeddings from human single-cell RNA profiles to aid cell-state and perturbation analysis.
What SE-600M is
SE-600M is a 600-million-parameter single-cell foundation model created by the Arc Institute. It forms part of the STATE model collection, designed to process human single-cell RNA expression profiles across biology, genomics, and medicine.
The model produces dense vector representations of individual cells, offering computational representations designed to assist researchers in evaluating diverse cell states and responses to biological perturbations.
Key facts
- Model collection
- STATE
- Parameter count
- 600M
- Data modality
- Human single-cell RNA expression
- Input format
- AnnData (.h5ad)
- Weight format
- SafeTensors
What you can use it for
- Cell-state characterisation
- Generates low-dimensional embeddings of human single-cell transcriptomes to help distinguish and evaluate cellular states.
- Perturbation analysis
- Provides single-cell representations to support the analysis and comparison of cellular responses under experimental perturbations.
- AnnData transcriptomic embedding
- Processes single-cell RNA expression matrices directly from AnnData (.h5ad) files to produce embedded AnnData outputs for downstream analysis.
Run it here
We list this model but cannot run it yet: no pipeline_tag — needs a human.
Everything on this page still applies. Running it needs a loader for this repository, not a bigger machine.
Strengths and limits
Good at
- Provides a specialised 600-million-parameter representation space for human single-cell transcriptomics.
- Operates directly on standard single-cell AnnData (.h5ad) structures via dedicated tooling.
Watch out for
- Targets human single-cell RNA expression profiles rather than non-human organisms.
- Requires execution through the dedicated arc-state tool and repository.
Inputs and outputs
You provide
Sequence or text
A protein or nucleotide sequence, or free text — whichever this model was trained on.
You get back
Mean-pooled representation, one vector per input.
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 SE-600M commercially?
Check the licence before commercial use. The See model card licence carries terms worth reading before you depend on it.
What kind of data does SE-600M process?
The model processes human single-cell RNA expression data supplied in AnnData (.h5ad) format.
What is the primary output generated by SE-600M?
It computes vector embeddings for each single-cell profile and writes them into an output AnnData (.h5ad) file.
Which software is required to run the model checkpoints?
The model checkpoints are designed to be run using the Arc State software package (arc-state).
Written from SE-600M'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.