Named Models
MedGemma 1.5 4BGoogleReleasedApr 2026Runs onnot in the notesClassmultimodal LLM
Update of MedGemma 4B (arXiv 2604.05081, 6 Apr 2026) adding 3D CT and MRI, whole-slide pathology, anatomical localization, longitudinal chest X-ray and document understanding. Reported gains over MedGemma 1 4B: MedQA +5%, EHRQA +22%, WSI macro F1 +47%, 3D MRI +11% (absolute or relative not stated).Commercial use: not stated; Jakarta hosting: unknown; Indonesian support: not stated
UNIMahmood Lab, HarvardReleasedMar 2024Runs onnot in the notesClassvisionPathology foundation model from the Mahmood Lab at Harvard, with weights for academic research only. Published in Nature Medicine on 19 Mar 2024.Commercial use: not stated; Jakarta hosting: unknown; Indonesian support: not stated
Prov-GigaPathMicrosoft AI, ProvidenceReleasednot in the notesRuns onnot in the notesClassvision
Pathology foundation model attributed to Microsoft and Providence; attribution unverified (page not read).Commercial use: not stated; Jakarta hosting: unknown; Indonesian support: not stated
VirchowPaigeReleasedJul 2024Runs onnot in the notesClassvisionPathology whole-slide image foundation model. Published in Nature Medicine on 22 Jul 2024 with pan-cancer specimen-level AUC 0.950.Commercial use: proprietary terms; Jakarta hosting: unknown; Indonesian support: not stated
CONCHMahmood Lab, HarvardReleasedJan 2024Runs onnot in the notesClassvisionMahmood Lab vision-language foundation model for computational pathology, trained on histopathology image-caption pairs.Commercial use: not stated; Jakarta hosting: unknown; Indonesian support: not stated
TITANMahmood Lab, HarvardReleasedNov 2024Runs onnot in the notesClassvision
Mahmood Lab multimodal whole-slide foundation model for pathology.Commercial use: not stated; Jakarta hosting: unknown; Indonesian support: not stated
UNI2-hMahmood Lab, HarvardReleasedJan 2025Runs onnot in the notesClassvisionSecond-generation UNI pathology foundation encoder from Mahmood Lab.Commercial use: not stated; Jakarta hosting: unknown; Indonesian support: not stated
Virchow2PaigeReleasedAug 2024Runs onnot in the notesClassvision
Paige pathology foundation model trained with mixed magnifications; successor of Virchow.Commercial use: not stated; Jakarta hosting: unknown; Indonesian support: not stated
H-optimus-0BioptimusReleasedJul 2024Runs onnot in the notesClassvisionPathology foundation vision transformer published by Bioptimus on Hugging Face.Commercial use: allowed; Jakarta hosting: unknown; Indonesian support: not stated
HibouHistAIReleasedJun 2024Runs onnot in the notesClassvision
Family of pathology foundation vision transformers from HistAI.Commercial use: allowed; Jakarta hosting: unknown; Indonesian support: not stated
PrismPaigeReleasedJul 2024Runs onnot in the notesClassmultimodal LLM
Pathology vision language model that writes reports from whole-slide image embeddings produced by Paige's Virchow encoder; research licence.Commercial use: non-commercial; Jakarta hosting: unknown; Indonesian support: not statedWhere this comes from
Models the research names for this use case. Research, with sources on each model record.
Who Runs It
No deployment is recorded for this use case.
Where this comes from
Documented uses by named organisations. The evidence column says who made the claim.
Candidate Models
Pan-cancer specimen-level detection (Virchow study)
A general benchmark for this model class, not a test on this use case. Specimen-level pan-cancer detection AUC on pathology slides. AUROC. Higher is better.
VirchowPaige vendor-reported0.95
UNIMahmood Lab, Harvard vendor-reported0.94
PhikonOwkin vendor-reported0.93CTransPath vendor-reported0.9100.51
RF-DETR 2XLRoboflow, visionReleasedNov 2025runs on edgehas scoreslast 18 months
RF-DETR MRoboflow, visionReleasedNov 2025runs on edgehas scoreslast 18 months
RF-DETR NRoboflow, visionReleasedNov 2025runs on edgehas scoreslast 18 months
RF-DETR SRoboflow, visionReleasedNov 2025runs on edgehas scoreslast 18 months
SAM 3Meta, visionReleasedNov 2025open weightshas scoreslast 18 monthsShowing 10 of 651. All models of this class
Where this comes from
Derived, not researched: models in the directory whose class matches a class this use case uses, excluding family records and models shut down or deprecated. Ranked by how many fit facts they meet, then by release date. Each fit fact is read from the model record.
Links
Models
VirchowPaige, vision
UNIMahmood Lab, Harvard, vision
Prov-GigaPathMicrosoft AI, vision
MedGemma 1.5 4BGoogle, multimodal LLM+7 more
VirchowPaige, vision
UNIMahmood Lab, Harvard, vision
Prov-GigaPathMicrosoft AI, vision
MedGemma 1.5 4BGoogle, multimodal LLM+7 moreWho Runs Itnone recorded
Models 11
MedGemma 1.5 4BGoogle, multimodal LLM
PrismPaige, multimodal LLM
CONCHMahmood Lab, Harvard, vision
Prov-GigaPathMicrosoft AI, Providence, vision
UNIMahmood Lab, Harvard, vision
UNI2-hMahmood Lab, Harvard, vision
VirchowPaige, vision
Virchow2Paige, vision
H-optimus-0Bioptimus, vision
HibouHistAI, vision
TITANMahmood Lab, Harvard, visionSources
- A foundation model for clinical-grade computational pathology and rare cancers detection - Nature Medicine nature.com https://www.nature.com/articles/s41591-024-03141-0
- Towards a general-purpose foundation model for computational pathology - Nature Medicine nature.com https://www.nature.com/articles/s41591-024-02857-3
- MedGemma 1.5 Technical Report arxiv.org https://arxiv.org/abs/2604.05081
- China Approves More AI Medical Devices, Expands Into Treatment letsdatascience.com https://letsdatascience.com/news/china-approves-more-ai-medical-devices-expands-into-treatmen-75de1963
To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Pathology Whole-Slide Imaging", https://aurion.semai.id/usecases/hc-pathology-wsi/, records of October 2026.