Bioptimus

Bioptimus describes an unnamed university IBD research team using H-Optimus-1 embeddings to predict spatial gene expression from H&E slides and prioritize slides/regions for spatial transcriptomics assays; evaluation compared 16 backbones with leave-one-patient-out validation.

Bioptimus reports usage by 16 of the top 20 pharma companies, more than 1,000 institutions and over one million model downloads, with models used across drug discovery, diagnostics and clinical research.

Bioptimus reports usage by 16 of the top 20 pharma companies, more than 1,000 institutions and over one million model downloads, with models used across drug discovery, diagnostics and clinical research.

Bioptimus describes H-Optimus-1 evaluation using STAMP on 575 training patients and 448 external-validation patients, with external risk-label classification AUROC 0.85.

Bioptimus introduced M-Optimus, a multimodal world model of biology spanning pathology, transcriptomics and genomics, and reports benchmark-leading H-Optimus models across hundreds of tasks.

Bioptimus introduced M-Optimus, a multimodal world model of biology spanning pathology, transcriptomics and genomics, and reports benchmark-leading H-Optimus models across hundreds of tasks.

Bioptimus announced reaching $76 million in cumulative funding, including a $41 million investment led by Cathay Innovation with participation from Sofinnova, Bpifrance and others.

Bioptimus is building STELA with partners including 10x Genomics and Broad Clinical Labs, targeting up to 100,000 patient specimens across three continents to support large-scale biological model development and deployment.

Bioptimus is building STELA with partners including 10x Genomics and Broad Clinical Labs, targeting up to 100,000 patient specimens across three continents to support large-scale biological model development and deployment.

Bioptimus's October 1, 2026 SDK release documents cached tissue masks, resumable cohort processing, stored run configurations, and Zarr/HDF5/NPZ outputs; the same pipeline supports embeddings and spatial-expression predictions by changing model arguments.

Bioptimus combines frontier ML with computational biology and clinical data science; the technical skills are highly valuable across AI-for-science employers but somewhat domain-specialized.

Bioptimus combines frontier ML with computational biology and clinical data science; the technical skills are highly valuable across AI-for-science employers but somewhat domain-specialized.