Cancer3.AIAI in OncologyAI Models › H-optimus-0
Foundation model Gated weights (licence click-through) Research use only pathology

H-optimus-0

1.1-billion-parameter ViT-g/14 pathology encoder trained on more than 500,000 H&E slides (hundreds of millions of tiles), released under Apache-2.0 — one of the few large pathology foundation models with a permissive licence.

80 579 downloads / month ♥ 92 updated 2025-12-12 gated

At a glance

DeveloperBioptimus
Released2024-07-10
Licenceapache-2.0
AvailabilityGated weights (licence click-through)
KindFoundation model
Regulatory statusResearch use only
Parameters1.1B

What it does

Tile embeddings for classification, mutation and biomarker prediction; the Apache licence makes it usable inside commercial pipelines.

Tasks, data types and cancers

CancerPan-cancer
Input224×224 tile at 20×
Output1,536-dim embedding

Architecture

FamilyVision Transformer
BackboneViT-g/14
Parameters1.1B
Pre-trainingself-supervised (DINOv2 family)

Training data

Over 500,000 H&E whole-slide images from a multi-site, multi-country collection assembled by Bioptimus; hundreds of millions of tiles.

Training set size>500,000 WSIs
InstitutionsBioptimus partner sites (undisclosed list)

Linked datasets

Evaluation

Benchmark / datasetMetricValueExternal validationSource
Camelyon16 — breast cancer detection (slide level, ABMIL) AUC-ROC 0.989 (SD 0.003); developer-run benchmark, average over 50 trainings no Source
PAIP 2020 — MSI prediction in colorectal cancer (slide level, ABMIL) AUC-ROC 0.967 (SD 0.007); developer-run benchmark, average over 50 trainings no Source
SLN-Breast — breast cancer detection (slide level, ABMIL) AUC-ROC 0.934 (SD 0.006); developer-run benchmark, average over 50 trainings no Source
TCGA-STAD-Kather — MSI prediction in gastric cancer (slide level, ABMIL) AUC-ROC 0.846 (SD 0.004); developer-run benchmark, average over 50 trainings no Source
Yale HER2 — HER2 status prediction in breast cancer (slide level, ABMIL) AUC-ROC 0.828 (SD 0.009); developer-run benchmark, average over 50 trainings no Source
CAMELYON17-WILDS, CRC-100K (normalised / not normalised), MHIST, TCGA-UNIFORM — tile classification (linear probing) accuracy 0.982 / 0.962 / 0.956 / 0.828 / 0.843 (SD 0.001-0.006); developer-run benchmark, linear classifier on frozen features, mean of three trainings no Source

How to run

# generic timm loader — check the model card for the exact init args and image normalisation
import timm, torch
from huggingface_hub import login
login()  # gated repos: accept the licence on huggingface.co first
model = timm.create_model('hf-hub:bioptimus/H-optimus-0', pretrained=True)
model.eval()
cfg = timm.data.resolve_data_config({}, model=model)
transform = timm.data.create_transform(**cfg)
# emb = model(transform(tile).unsqueeze(0))  # 1 x D tile embedding
Librarytimm

Gated download (accept terms); the model card specifies the exact normalisation constants — use them, not ImageNet defaults.

Regulatory status and intended use

Regulatory statusResearch use only
Intended useResearch; not a medical device.

Source →

Regulatory status is quoted from the source linked above and can change. Research-use-only models must not be used for clinical decisions.

Limitations and bias

  • No peer-reviewed paper at release; details of the pretraining cohort are only partially disclosed.
  • Large model: slower tile throughput than ViT-L encoders.

Sources

  1. Hugging Face model card — bioptimus/H-optimus-0
  2. Bioptimus releases (GitHub) — H-optimus-0 evaluation protocol and results tables

This page is educational — it is not medical advice and does not replace consultation with an oncologist. Diagnostic and treatment decisions are made solely by specialist physicians.