Foundations
How we evaluate AI models for healthcare (our editorial standards)
The editorial standards behind the OpenMed Router blog: how we assess open source models, verify compliance claims, and keep healthcare teams up-to-date with accurate, trustworthy information. Written by Chris Williams, MD.
TL;DR
How we assess open source models, verify compliance claims, and keep health-tech teams up-to-date with accurate, trustworthy information.
Healthcare teams adopting AI face a real problem: most of the information about models online is either marketing from the company selling the model or generic listicles with no clinical or compliance lens. This blog exists to fix that. This short post explains how we evaluate models and write about them, so you can judge whether to trust what you read here.
Who writes this
I'm Chris Williams, MD. I'm a physician and the founder of OpenMed Router, where we're building HIPAA-compliant cloud inference for open source models. I write here about clinical AI, model selection, compliance, and what it actually takes to put an open source model into a healthcare workflow without breaking the rules. When a post touches a domain I'm not expert in, I have it reviewed by a colleague before publishing.
Our editorial standards
Accuracy over speed. A model spec, a price, or a compliance claim gets verified before it goes in a post. We would rather publish late and be right than news-jack and be wrong.
Honest about the gaps. AI models are only as good as the accuracy of the claims made about them. We verify specs, compliance status, and clinical appropriateness before publishing.
Open source, not advocacy. Open source models are often cheaper and more flexible than closed-source frontier models, but they are not always the right answer. We compare honestly, including where a managed closed-source API is the better choice for a given workload.
Primary sources first. Specs and pricing come from the model provider's published documentation. Compliance status comes from the actual BAA terms and the hosting provider's attestations, not from a blog post about them.
What we cover
- Compliance: what HIPAA, BAAs, and related frameworks require for AI inference.
- Model launches: what a new open source model is, its specs, and what it means for healthcare workloads.
- Comparisons: how OpenAI, Anthropic, Azure OpenAI, and AWS Bedrock compare against open source inference for health-tech teams.
- Foundations: plain-English definitions for the AI terms health-tech builders need to know.
How you can help
If you spot something inaccurate in a post, or a spec that has changed, email hello@openmedrouter.com. We will correct it and note the change.
Chris Williams, MD
Chris Williams, MD is a physician, technologist and the co-founder of OpenMed Router, working to make open source AI models safely accessible to healthcare organizations under HIPAA. He writes about clinical AI, model selection, compliance, and the practical adoption of open source inference in clinical and operational workflows.
Join the waitlist