{"slug":"foundation-model","tag":"foundation-model","variants":["foundation-model"],"description":"Foundation models trained on pathology, radiology or molecular data, and the groups behind them.","count":43,"kinds":{"technology":42,"company":1},"related":[{"slug":"pathology","tag":"pathology","shared":13},{"slug":"virtual-cell","tag":"virtual-cell","shared":10},{"slug":"genome","tag":"genome","shared":5},{"slug":"radiology","tag":"radiology","shared":5},{"slug":"protein-design","tag":"protein-design","shared":3},{"slug":"structure","tag":"structure","shared":3},{"slug":"ehr","tag":"ehr","shared":1},{"slug":"llm","tag":"llm","shared":1},{"slug":"multimodal","tag":"multimodal","shared":1},{"slug":"phenomics","tag":"phenomics","shared":1}],"records":[{"id":"virchow","kind":"technology","name":"Virchow / Virchow2 (Paige, MSK)","route":"/technologies/virchow/","status":"emerging","tldr":"A pathology foundation model trained on millions of slides that can detect cancer and predict biomarkers from an ordinary H&E slide."},{"id":"uni-conch","kind":"technology","name":"UNI and CONCH (Harvard, Mahmood Lab)","route":"/technologies/uni-conch/","status":"emerging","tldr":"Two open academic pathology models: UNI reads tissue images, CONCH links images with pathology text."},{"id":"titan","kind":"technology","name":"TITAN (whole-slide multimodal model)","route":"/technologies/titan/","status":"emerging","tldr":"TITAN is a model that summarises a whole slide, not just tiles, and can write a draft pathology report."},{"id":"prov-gigapath","kind":"technology","name":"Prov-GigaPath (Microsoft, Providence)","route":"/technologies/prov-gigapath/","status":"emerging","tldr":"An open pathology model trained on 1.3 billion image tiles from a US health system, modelling whole slides at gigapixel scale."},{"id":"chief","kind":"technology","name":"CHIEF (Harvard, Yu Lab)","route":"/technologies/chief/","status":"emerging","tldr":"A pathology model trained across 19 cancer types that predicts survival and mutations from slides."},{"id":"musk","kind":"technology","name":"MUSK (Stanford, vision-language pathology)","route":"/technologies/musk/","status":"emerging","tldr":"A model that reads slides and clinical text together to predict who will respond to immunotherapy."},{"id":"h-optimus","kind":"technology","name":"H-optimus (Bioptimus)","route":"/technologies/h-optimus/","status":"emerging","tldr":"An open 1.1-billion-parameter pathology model from a French startup, among the strongest on public benchmarks."},{"id":"pluto","kind":"technology","name":"PLUTO (PathAI)","route":"/technologies/pluto/","status":"emerging","tldr":"PLUTO is PathAI's compact pathology foundation model, a vision transformer pretrained at several magnifications on 195 million tiles from 158,000 slides, so one network serves slide-level and biomarker quantification tasks at whatever resolution each needs. It runs inside PathAI's AISight product, but its weights are proprietary, so outside groups cannot benchmark or adapt it."},{"id":"hibou","kind":"technology","name":"Hibou (HistAI)","route":"/technologies/hibou/","status":"emerging","tldr":"Hibou is a family of open pathology foundation models under a permissive licence."},{"id":"kaiko-midnight","kind":"technology","name":"Midnight (kaiko.ai)","route":"/technologies/kaiko-midnight/","status":"emerging","tldr":"Midnight is a pathology model that matched the leaders while training on far fewer slides."},{"id":"atlas-aignostics","kind":"technology","name":"Atlas (Aignostics, Mayo Clinic, Charité)","route":"/technologies/atlas-aignostics/","status":"emerging","tldr":"Atlas is a pathology foundation model trained on 1.2 million slides from two of the world's largest hospitals."},{"id":"phikon","kind":"technology","name":"Phikon / Phikon-v2 (Owkin)","route":"/technologies/phikon/","status":"emerging","tldr":"Owkin's open pathology models trained on TCGA and its federated hospital network."},{"id":"merlin-ct","kind":"technology","name":"Merlin (Stanford abdominal CT vision-language model)","route":"/technologies/merlin-ct/","status":"emerging","tldr":"Merlin is a model trained on 15,000 CT scans with their reports that can find and describe hundreds of findings."},{"id":"radfm","kind":"technology","name":"RadFM (generalist radiology foundation model)","route":"/technologies/radfm/","status":"emerging","tldr":"An open generalist model that answers questions about 2D and 3D scans."},{"id":"ct-fm","kind":"technology","name":"CT-FM (whole-body CT foundation model)","route":"/technologies/ct-fm/","status":"emerging","tldr":"A model pretrained on 148,000 CT scans to segment organs and triage findings."},{"id":"medsam","kind":"technology","name":"MedSAM / SAM-Med3D (segment anything for medicine)","route":"/technologies/medsam/","status":"emerging","tldr":"Adaptations of Meta's Segment Anything model that outline tumours and organs on any scan with a click."},{"id":"med-gemini","kind":"technology","name":"Med-Gemini and MedLM (Google)","route":"/technologies/med-gemini/","status":"emerging","tldr":"Google's medical versions of its Gemini models, able to reason over text, images, and long records."},{"id":"aidoc-care","kind":"technology","name":"Aidoc CARE (clinical radiology foundation model)","route":"/technologies/aidoc-care/","status":"emerging","tldr":"Aidoc CARE is one radiology foundation model, pretrained on CT scans without labels, whose task-specific heads have each been FDA-cleared to flag urgent findings in emergency scans so radiologists read those first. Its oncology relevance is indirect, catching incidental masses; the regulatory evidence covers triage, not diagnostic accuracy for tumours."},{"id":"geneformer","kind":"technology","name":"Geneformer","route":"/technologies/geneformer/","status":"emerging","tldr":"Geneformer is a transformer trained on about 30 million single cells that encodes each cell as a ranked list of its genes, so deleting a gene in silico shows which genes matter in a disease. It was the first single-cell foundation model in general use, though benchmarks find only modest gains over linear baselines on some tasks."},{"id":"scgpt","kind":"technology","name":"scGPT","route":"/technologies/scgpt/","status":"emerging","tldr":"A GPT-style model for single-cell data that predicts cell types, perturbation responses, and gene networks."},{"id":"universal-cell-embedding","kind":"technology","name":"Universal Cell Embedding (UCE)","route":"/technologies/universal-cell-embedding/","status":"emerging","tldr":"Universal Cell Embedding maps any cell from any species into one shared space without retraining."},{"id":"scfoundation","kind":"technology","name":"scFoundation (BioMap)","route":"/technologies/scfoundation/","status":"emerging","tldr":"scFoundation is a 100-million-parameter model trained on 50 million cells, from China's BioMap."},{"id":"nicheformer","kind":"technology","name":"Nicheformer (spatial single-cell)","route":"/technologies/nicheformer/","status":"emerging","tldr":"Nicheformer is a model trained on both dissociated and spatial data so it learns how a cell's neighbourhood shapes it."},{"id":"cellfm","kind":"technology","name":"CellFM","route":"/technologies/cellfm/","status":"emerging","tldr":"CellFM is an 800-million-parameter single-cell model trained on 100 million human cells."},{"id":"genept","kind":"technology","name":"GenePT","route":"/technologies/genept/","status":"emerging","tldr":"Uses text embeddings of gene descriptions from a general LLM to represent cells, and performs surprisingly well."},{"id":"c2s-scale","kind":"technology","name":"Cell2Sentence / C2S-Scale (Yale, Google)","route":"/technologies/c2s-scale/","status":"emerging","tldr":"Turns a cell's gene expression into a sentence so a normal language model can reason about it; a 27-billion-parameter version proposed a cancer immunotherapy idea that was confirmed in the lab."},{"id":"state-arc","kind":"technology","name":"State (Arc Institute perturbation model)","route":"/technologies/state-arc/","status":"emerging","tldr":"Predicts how cells will respond to a drug or gene knockout, trained on over 100 million perturbed cells."},{"id":"transcriptformer","kind":"technology","name":"TranscriptFormer and rBio (CZI virtual cell models)","route":"/technologies/transcriptformer/","status":"emerging","tldr":"CZI's open cross-species cell models and a reasoning model trained on them."},{"id":"alphafold3","kind":"technology","name":"AlphaFold 3","route":"/technologies/alphafold3/","status":"established","tldr":"Predicts the 3D shape of proteins together with DNA, RNA, small molecules and antibodies, the starting point for much modern drug design."},{"id":"boltz","kind":"technology","name":"Boltz-1 / Boltz-2 (MIT, open)","route":"/technologies/boltz/","status":"emerging","tldr":"Open-source structure models that match AlphaFold 3, with Boltz-2 also predicting how strongly a drug binds."},{"id":"chai-1","kind":"technology","name":"Chai-1 / Chai-2","route":"/technologies/chai-1/","status":"emerging","tldr":"Structure and antibody-design models from Chai Discovery, with Chai-2 reporting high zero-shot antibody hit rates."},{"id":"esm3","kind":"technology","name":"ESM3 (EvolutionaryScale)","route":"/technologies/esm3/","status":"emerging","tldr":"ESM3 is a generative protein model that designed a working fluorescent protein far from any natural sequence."},{"id":"evo2","kind":"technology","name":"Evo 2 (Arc Institute, NVIDIA)","route":"/technologies/evo2/","status":"emerging","tldr":"A DNA language model trained on 9.3 trillion bases that can flag cancer-causing BRCA1 variants without being told about them."},{"id":"rfdiffusion","kind":"technology","name":"RFdiffusion / RFdiffusion2 and ProteinMPNN (Baker Lab)","route":"/technologies/rfdiffusion/","status":"emerging","tldr":"The tools that design entirely new proteins to bind a chosen target, now used for cancer binders and antibodies."},{"id":"bioemu","kind":"technology","name":"BioEmu (Microsoft)","route":"/technologies/bioemu/","status":"emerging","tldr":"BioEmu is a Microsoft generative diffusion model that predicts the range of shapes a protein moves between, not one static structure, thousands of times faster than molecular dynamics simulation. For cancer drug discovery that can reveal transient pockets, as in KRAS, that static predictors miss, but its outputs are approximate and validated mainly on small proteins."},{"id":"nucleotide-transformer","kind":"technology","name":"Nucleotide Transformer (InstaDeep)","route":"/technologies/nucleotide-transformer/","status":"emerging","tldr":"DNA language models trained on thousands of genomes for variant and regulatory prediction."},{"id":"enformer-borzoi","kind":"technology","name":"Enformer and Borzoi (DeepMind, Calico)","route":"/technologies/enformer-borzoi/","status":"emerging","tldr":"Models that predict how DNA sequence controls gene activity, used to interpret non-coding cancer mutations."},{"id":"alphamissense","kind":"technology","name":"AlphaMissense","route":"/technologies/alphamissense/","status":"established","tldr":"Scored all 71 million possible single-letter protein changes in humans as likely harmful or benign."},{"id":"alphagenome","kind":"technology","name":"AlphaGenome","route":"/technologies/alphagenome/","status":"emerging","tldr":"Reads a million letters of DNA at once and predicts how a mutation changes gene regulation, splicing and chromatin."},{"id":"foresight-ehr","kind":"technology","name":"Foresight (generative EHR model)","route":"/technologies/foresight-ehr/","status":"emerging","tldr":"A model trained on millions of hospital records that forecasts a patient's next diagnoses."},{"id":"tempus-multimodal","kind":"technology","name":"Tempus multimodal models","route":"/technologies/tempus-multimodal/","status":"emerging","tldr":"Models trained on Tempus's paired genomic, pathology, imaging and outcome data to predict response and prognosis."},{"id":"phenom-2","kind":"technology","name":"Phenom-2 and Recursion OS","route":"/technologies/phenom-2/","status":"emerging","tldr":"A model trained on billions of cell microscopy images to read what a drug or gene knockout does to a cell."},{"id":"strand-ai","kind":"company","name":"Strand AI","route":"/companies/strand-ai/","tldr":"Strand AI's first model, Lattice, predicts which proteins sit where across a tumour from the ordinary stained slide a hospital already has, so a whole archive can be profiled without spending tissue on extra laboratory staining."}]}