BioMap
Run the xTrimoPGLM protein model and xTrimoGene single-cell model through BioMap's API for large-scale protein and transcriptomics tasks.
Overview
BioMap develops the xTrimo family of large-scale biological foundation models and serves them through its own API. Teams that want to run high-capacity protein and single-cell models — checkpoints too large to run casually on local hardware — can access xTrimoPGLM and xTrimoGene through BioMap without provisioning the infrastructure themselves.
What you can run on BioMap
- xTrimoPGLM — a unified 100-billion-parameter protein language model that combines autoencoding and autoregressive objectives in a single backbone, spanning protein understanding and generation. It produces embeddings and predictions across protein structure, function, and property tasks.
- xTrimoGene — an asymmetric encoder-decoder transformer for single-cell RNA-seq data that cuts compute by one to two orders of magnitude, used for cell representation and transcriptomics downstream tasks.
These span large-scale protein language modeling and single-cell transcriptomics. The models are served through BioMap's API rather than distributed as downloadable weights.
API access on BioMap
BioMap's API is the primary way to run xTrimoPGLM and xTrimoGene, giving teams hosted inference on models whose scale makes local deployment impractical. This is aimed at protein scientists, computational biologists, and single-cell researchers who want access to very large protein and transcriptomics foundation models on demand. Because these are served rather than distributed, the API is the route to running them at full scale, with request-based access documented by BioMap.
Run inference on BioMap (2)
Unified 100-billion-parameter protein language model combining autoencoding and autoregressive objectives for protein understanding and generation.
Asymmetric encoder-decoder transformer for single-cell RNA-seq that encodes only non-zero genes, cutting FLOPs 10-100x versus standard transformers.