Multi-Modal Foundation · living-models
Botanic1-S
A 318M-parameter bidirectional Mamba-2 foundation model for analysing plant genomes at single-nucleotide resolution.
What Botanic1-S is
Botanic1-S is a genomic foundation model developed by living-models for plant biology and genomics research. It provides representations of plant DNA sequences to support downstream genomic analyses across diverse plant taxa.
Plant genomes often present unique structural characteristics that require specialised computational approaches. Pre-trained across 320 embryophyte (land plant) species, Botanic1-S provides nucleotide-level sequence representations to aid the study of plant genetics.
How it works
Botanic1-S is based on a bidirectional Mamba-2 encoder architecture containing 318 million parameters. It processes genomic sequences with an 8,192-token context window at single-nucleotide resolution.
The model was pre-trained on a corpus of 314.6 billion tokens gathered from 320 embryophyte species.
Key facts
- Architecture
- Bidirectional Mamba-2 encoder
- Parameter count
- 318M
- Context length
- 8,192 tokens
- Input resolution
- Single-nucleotide resolution
- Pre-training data volume
- 314.6B tokens
- Species coverage
- 320 embryophyte species
What you can use it for
- Plant genomic sequence analysis
- Extract representations of plant DNA sequences across context windows of up to 8,192 tokens.
- Single-nucleotide resolution modelling
- Evaluate sequence features and nucleotide-level variation in plant genomes.
- Comparative embryophyte genomics
- Study sequence patterns across diverse land plant species using representations trained on 320 embryophyte taxa.
Run it here
We list this model but cannot run it yet: gated on the Hub — never run past a gate the user did not accept.
Everything on this page still applies. Running it needs a loader for this repository, not a bigger machine.
Strengths and limits
Good at
- Pre-trained on 314.6 billion tokens spanning 320 embryophyte species.
- Processes genomic inputs at single-nucleotide resolution.
- Employs a bidirectional Mamba-2 encoder architecture that scales to an 8,192-token context length.
Watch out for
- Sequence inputs are restricted to an 8,192-token context length.
- Requires custom code execution to load and run the model architecture.
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 · other
This model is gated by its authors — its licence has to be accepted on the source repository before it will run.
Questions
Can I use Botanic1-S commercially?
Check the licence before commercial use. This model is gated by its authors — its licence has to be accepted on the source repository before it will run.
What architecture is used by Botanic1-S?
Botanic1-S is a 318M-parameter bidirectional Mamba-2 encoder.
What is the context window of the model?
The model has an 8,192-token context length at single-nucleotide resolution.
Which organisms were included in the training data?
The pre-training dataset consists of 314.6 billion tokens derived from 320 embryophyte (land plant) species.
Who developed Botanic1-S?
The model was developed by living-models.
Written from Botanic1-S'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.