PAICON-GmbH/Athena-1

Athena-1: PAICON's vision foundation model for computational pathology

Built on a customized DINOv3 self-supervised stack, Athena-1 delivers strong, broad transfer across tissue classification and spatial gene-expression tasks.

Model Details

• Model type: ViT-G/14
• Params: ~1.1B
• Input: RGB patches 224×224
• Output: 1536-dim features (CLS)

Training Data

• Slides: ~282,500 H&E WSIs
• Patches: ~115 million
• Diversity: Multi-country (25), multi-institution, 8 scanner models, broad organ coverage

Powered by PAICON's PAIX tech stack.

Part of the metadata and Whole Slide Images(WSI) used for Athena-1 training are available through: https://paix-navigator.paicon.com/beta_access_requests/new

Benchmark

Model	HEST	BACH	BreakHis	CRC	Gleason	MHIST	PCam	PCam/test
Athena-1	0.420	0.922	0.809	0.970	0.769	0.839	0.939	0.952
Athena-0	0.387	0.865	0.789	0.970	0.740	0.852	0.944	0.951
Virchow2	0.398	0.883	0.821	0.967	0.783	0.861	0.933	0.938
UNI2	0.414	0.915	0.859	0.965	0.775	0.824	0.944	0.950
H-optimus-0	0.415	0.759	0.801	0.955	0.770	0.843	0.932	0.943

Usage

from transformers import AutoModel
from torchvision import transforms
import torch
from PIL import Image

model = AutoModel.from_pretrained("PAICON-GmbH/Athena-1", trust_remote_code=True) 
model.eval()

transform = transforms.Compose([
    transforms.Resize((224, 224)),
    transforms.ToTensor(),
    transforms.Normalize(
        mean=(0.707223, 0.578729, 0.703617),
        std=(0.211883, 0.230117, 0.177517),
    ),
])

image = Image.open("tile.png").convert("RGB")
x = transform(image).unsqueeze(0)

with torch.no_grad():
    cls_features = model(x)                                       # [1, 1536]
    cls_features, patch_features = model.forward_with_patches(x)  # [1, 1536], [1, 256, 1536]