AI Capability
Which healthcare AI vendors disclose what is under the hood?
The AI Health Index grades all 670 vendors on Model and Technology Transparency, one of 15 capability axes applied to every record without exception. 21 of 670 vendors grade A, 360 grade B, 282 grade C and 7 grade D. That places this axis 11th of 15 by the number of vendors reaching the top grade. Grades were last verified on October 20, 2026 and are never aggregated into a composite score.
What this axis measures
Disclosure of what is under the hood: proprietary models versus fine-tuned foundation models, training data claims, model cards, versioning and update practices.
Buyers also search this as: what model does it use, proprietary versus foundation models, training data, model versioning, and AI transparency.
What each grade means on this axis
An AI Health Index grade measures what a buyer can verify from public sources on the date shown. It is not a rating of how good the product is. A vendor can build an excellent system and grade low on an axis because it publishes nothing an outsider can check.
The distribution
Reading the result
Proprietary AI is the most overloaded phrase in this market. It covers genuinely trained models, fine tunes of open weights, and prompt engineering over a commercial API, and a buyer cannot tell which from the marketing page of most vendors in the index.
The reason it matters is not intellectual curiosity. It determines who else is in the data path, how the product behaves when an upstream model changes underneath it, and whether the vendor can tell you that a change happened at all. Versioning practice is the tell: a vendor that publishes what changed and when is a vendor that knows.
Where the A grades are, by category
Categories are shown by the share of their vendors reaching an A. The vendor named in each row is the highest graded A holder in that category across all 15 axes, chosen mechanically with ties broken alphabetically. Categories with no A holder on this axis are omitted.
| Category | A grades | Share | Leading vendor |
|---|---|---|---|
| Inpatient Deterioration & Risk Monitoring | 3 of 25 | 12% | AgileMD |
| Drug Discovery AI | 6 of 61 | 10% | EvolutionaryScale |
| Clinical Decision Support | 11 of 143 | 8% | Cytovale IntelliSep |
| Diagnostics & Genomics | 5 of 87 | 6% | Cytovale IntelliSep |
| Radiology & Imaging AI | 3 of 53 | 6% | Lunit |
| Clinical Trials AI | 2 of 52 | 4% | Anumana |
| Medication Safety & Prescribing | 1 of 23 | 4% | InsightRX |
| Behavioral Health AI | 1 of 29 | 3% | Canary Speech |
| Digital Pathology AI | 1 of 37 | 3% | ArteraAI |
| Ambient Scribes | 2 of 124 | 2% | TORTUS |
| Patient Voice Agents | 1 of 59 | 2% | Hippocratic AI |
| Remote Monitoring & Chronic Care | 1 of 44 | 2% | Canary Speech |
Questions worth asking a vendor
- Is the model proprietary, a fine tune, or an API call, and which parts are which?
- What happens to product behavior when an upstream model updates, and are we told?
- Is there a version history a buyer can read?
Questions buyers ask
How many healthcare AI vendors grade well on model and technology transparency?
Of the 670 vendors in the AI Health Index, 21 grade A on this axis, 360 grade B, 282 grade C and 7 grade D under the AI Health Index grading framework. Grades were last verified on October 20, 2026. Grades are not aggregated into a composite score.
What does an A grade mean on model and technology transparency?
Disclosure of what is under the hood: proprietary models versus fine-tuned foundation models, training data claims, model cards, versioning and update practices. What is under the hood is named: proprietary or adapted foundation models identified, training data characterized, and versioning and update practice published so a buyer knows when the system changed.
What does a D grade mean on model and technology transparency?
Nothing is published about what produces the output. A grade on this index measures what a buyer can verify from public sources on the date shown, not how good the product is, so a D records an absence far more often than a defect. A vendor that publishes more is regraded.
Do vendors pay to be included or graded?
No. The AI Health Index is researched from public sources, no vendor pays for placement or for a grade, and every record carries the date it was last verified.
The other 14 axes
No single axis decides a selection. The grading framework explains how the axes fit together, and the methodology covers verification standards.