When is Aidoc the better choice?

The foundation model approach produced a clearance covering many acute indications at once, with regulator reviewed performance figures rather than a vendor benchmark behind them. The operating system handles normalisation, continuous performance monitoring and governance across every model it runs, which is the part a health system cannot build itself and the reason multi algorithm deployments usually fail. The data handling model is published in unusual detail, with de identification performed before upload, which changes the health privacy question architecturally rather than contractually.

When is Subtle Medical the better choice?

It does not detect anything, which is the point: enhancement raises image quality or cuts scan time on scanners the department already owns, so the benefit is throughput rather than a new finding. A reported eleven clearances across two modalities with more than 25 peer reviewed publications, and a published standards based integration specification. The economics are different in kind. Detection AI adds work to the read; enhancement takes time off the scanner, which is the constraint most imaging departments are actually managing.

Aidoc vs Subtle Medical

Last VerifiedAugust 4, 2026

Verdict

Two of the strongest records in imaging AI doing entirely different jobs, which is worth stating because they compete for the same budget. Aidoc finds things: an operating system running detection models across an imaging estate, with a January 2026 clearance spanning many acute indications from one named foundation model and regulator reviewed performance figures behind it. Subtle changes how the image looks: deep learning reconstruction and denoising that lets a department scan faster or at lower dose on the scanners it already owns, with eleven clearances and more than 25 publications. The budget question is which constraint binds. If studies are read too late, Aidoc. If the scanner is the bottleneck and the waiting list is the problem, Subtle returns capacity without adding anything to the radiologist's read, and that is a different kind of win.

Select Aidoc if

The foundation model approach produced a clearance covering many acute indications at once, with regulator reviewed performance figures rather than a vendor benchmark behind them.
The operating system handles normalisation, continuous performance monitoring and governance across every model it runs, which is the part a health system cannot build itself and the reason multi algorithm deployments usually fail.
The data handling model is published in unusual detail, with de identification performed before upload, which changes the health privacy question architecturally rather than contractually.

Select Subtle Medical if

It does not detect anything, which is the point: enhancement raises image quality or cuts scan time on scanners the department already owns, so the benefit is throughput rather than a new finding.
A reported eleven clearances across two modalities with more than 25 peer reviewed publications, and a published standards based integration specification.
The economics are different in kind. Detection AI adds work to the read; enhancement takes time off the scanner, which is the constraint most imaging departments are actually managing.

Attribute Matrix

Side by Side

Axis	A Aidoc	S Subtle Medical
AI Centrality
Autonomy and Oversight Model
Model and Technology Transparency
Clinical and Operational Evidence
AI Safety and PHI Stewardship
HIPAA and BAA Posture
Security Certifications and Trust Center
FDA and Regulatory Status
AI Governance and Bias Disclosure
EHR and Interoperability Depth
Deployment Model and Data Residency
Commercial Transparency
Setting and Specialty Coverage

Keep Comparing

Related comparisons

Other published head to head assessments involving these vendors or their closest peers. The full set for this category is on the Radiology & Imaging AI page.

Aidoc vs Bunkerhill Health→Aidoc vs Qure AI→Aidoc vs Rad AI→Aidoc vs Viz.ai→AIRS Medical vs Subtle Medical→

Disclosure

These products do different jobs and a department may well buy both, so the comparison is about which constraint binds rather than which vendor wins. Their failure modes are opposite and both are easy to miss: a triage model that degrades quietly returns worklists to their usual order, while a reconstruction model that degrades produces cleaner images that have lost something, and only the second is invisible to the reader. Neither vendor publishes pricing.

Subtle publishes no subgroup performance and its attestation is under specified; Aidoc claims its own certification with an annually certified management system, which is the stronger position of the two.