Dyania Health vs Mendel (2026)
Dyania Health builds Synapsis AI, a medical language model that reads a health system's full records and answers set clinical questions, such as whether a patient meets each criterion of a trial. It is installed on premise or in the health system's private cloud. Dyania's own physicians check positives and negatives before results are delivered. Health systems buy it for trial prescreening, registries and studies, and Cleveland Clinic has rolled it out across its system. Mendel sells Hypercube, a platform that reads unstructured records for chart review, cohort building and trial prescreening. It pairs language models with symbolic reasoning over Mendel's own clinical hypergraph and links every answer to highlighted text in the record. It can run in the customer's cloud or be hosted by Mendel, and it connects to warehouses such as Databricks and Snowflake. Life sciences and diagnostics companies buy it, as do health systems. What separates them is who does the checking. Dyania's physicians review the model's work before the customer sees it, so a health system receives confirmed candidate lists. Mendel's customers check the work themselves, with the source evidence highlighted beside each answer, inside their own data stack. A health system that wants vetted trial candidates delivered fits Dyania. A research or data team that wants to run abstraction and cohort work itself fits Mendel.
- Dyania's physicians review true and false positives and negatives before any result reaches the customer.
- Synapsis AI runs on premise or in the health system's private cloud, and records do not leave it.
- At Cleveland Clinic it reviewed 1,476 patients in a week for a heart failure trial, and 29 of its 30 matches had been missed by usual screening.
- In a polycythemia vera trial search, all 22 patients it flagged were confirmed eligible.
- Study feasibility, registry reporting, observational studies and care pathways run on the same model.
- Every answer links to highlighted evidence in the original record.
- Hypercube connects to Databricks and Snowflake and to business intelligence tools such as Tableau and Qlik.
- It can run inside the customer's own cloud, or Mendel can host it.
- In a randomized University of Pennsylvania study, staff using it reached 76.5 percent accuracy on trial eligibility, against 71.1 percent without it.
- Chart review, cohort analysis and trial eligibility mapping are sold as separate configurations.
This comparison is published by AI Health Index, an independent research platform that compares healthcare AI vendors objectively. Dyania Health and Mendel are each graded against the same capability taxonomy, from each vendor's own public materials and the regulatory record, under the AI Health Index verification standard. No vendor pays for placement, and no vendor has reviewed this page. How this evidence is graded
Plain facts
| Fact | Dyania Health | Mendel |
|---|---|---|
| Primary category | Clinical Trials AI | Clinical Trials AI |
| Founded | 2019 | 2016 |
| Headquarters | Jersey City, New Jersey, United States | San Jose, California, United States |
| Website | dyaniahealth.com | mendel.ai |
Side by Side
Select any grade to read the note behind it.
Each vendor in one paragraph
Written to be quoted whole. Each paragraph states what the AI Health Index verified about the vendor, with the caveats attached.
Dyania Health sells Synapsis AI, a medical language model that reads full electronic medical records and answers specific clinical questions, to health systems and drug sponsors for trial prescreening, study feasibility, registries and observational research. It runs on premise or in the health system's private cloud, and Dyania's physicians review positive and negative results before delivery. At Cleveland Clinic, which has rolled it out system wide and invested in the company, a 2026 Journal of Cardiac Failure study reported 96.2 percent accuracy and found 29 trial matches that usual screening had missed. The full paper sits behind a paywall, no security certification is published, and Dyania publishes no price.
Source: AI Health Index, October 2026
Mendel sells Hypercube, a platform that reads unstructured patient records for chart review, cohort building and trial prescreening, to life sciences and diagnostics companies and to health systems. It pairs large language models with symbolic reasoning over a clinical hypergraph, links each answer to highlighted evidence in the record, and can run in the customer's own cloud or be hosted by Mendel. A randomized University of Pennsylvania and Emory evaluation in Nature Communications (2026) found staff using the AI 76.5 percent accurate against 71.1 percent without it, with no time saved, and the AI alone at 59.9 percent. Mendel publishes no security certification and no rate card on its own site.
Source: AI Health Index, October 2026
Questions buyers ask
How does Dyania Health compare to Mendel?
Dyania Health and Mendel both use AI to read unstructured patient records for trial prescreening and research. Dyania installs its Synapsis AI model inside the health system and has its own physicians check the results before they are delivered. Mendel sells Hypercube as a platform the customer's own team runs, with every answer linked to highlighted evidence in the record and connections to data warehouses such as Snowflake and Databricks.
Dyania suits a health system that wants vetted candidate lists, and Mendel suits a data or research team that wants to run the work itself. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
How much do Dyania Health and Mendel cost?
Dyania Health publishes no price. It says health systems of about 500,000 patients spend more than $5 million a year on manual chart abstraction, and positions savings of up to 50 percent against that. Mendel publishes no rate card on its own site. Its AWS Marketplace listing sells Hypercube on contract with one dimension of $1,000,000 a month per instance, and says actual pricing depends on the contract agreed with Mendel. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
What evidence supports Dyania Health and Mendel?
Mendel's evidence is a randomized evaluation led by the University of Pennsylvania and Emory and published in Nature Communications in 2026. Across 355 lung and colorectal cancer charts, research staff using Mendel's AI reached 76.5 percent chart level accuracy on 12 eligibility criteria against 71.1 percent without it, and saved no time; the AI alone scored 59.9 percent, and seven of the 17 authors work for Mendel.
Dyania Health's evidence comes from Cleveland Clinic, which has invested in the company. A 2026 Journal of Cardiac Failure study there reported 96.2 percent accuracy across 7,700 questions, and a 2025 hematology study found all 22 flagged patients eligible. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
Where do Dyania Health and Mendel run, and what do they connect to?
Dyania Health installs Synapsis AI on premise or in the health system's private cloud, reads the record where it sits, and names no EHR product or integration standard; its staff reach a sandboxed environment through the health system's VPN. Mendel's Hypercube can run inside the customer's own cloud, where data stays within the customer's walls, or on Mendel's own hosting on AWS, and it connects to Databricks, Snowflake, Tableau and Qlik rather than to a named EHR.
Neither vendor publishes a SOC 2 report, HITRUST certification, ISO 27001 certificate or business associate agreement. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
Are Dyania Health and Mendel regulated by the FDA?
Neither Dyania Health nor Mendel claims FDA clearance, and both scope their products to chart review, research and trial prescreening that people confirm. The University of Pennsylvania's trial registrations record Mendel's product as neither an FDA regulated drug nor a device. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
Related comparisons
Other published head to head assessments involving these vendors or their closest peers. The full set for this category is on the Clinical Decision Support page.
Dyania Health is based in Jersey City with an office in Athens. Cleveland Clinic, where its published studies were run, has invested in the company. Mendel is based in San Jose and backed by Oak HC/FT and DCM, and seven of the 17 authors on its Penn evaluation work for Mendel. Neither vendor publishes a security certification, a business associate agreement or a rate card on its own site. Dyania has not said how many eligible patients its model misses or whose records trained it. Mendel's Penn evaluation found staff sometimes followed wrong AI answers on performance status, and how the product guards against that is not described.