You look something up between patients, get your answer, and move on. By tonight you have forgotten it. That gap between finding a fact and actually learning it is the one OpenEvidence says its new Learn platform is built to close, and it is free.
On November 4, 2026, OpenEvidence announced Learn, a platform for medical students, residents and practicing clinicians that bundles a question bank, an AI tutor, flashcards and CME credit. Here’s the thing: OpenEvidence started as a tool for answering clinical questions at the point of care. Learn tries to turn those same questions into study material.
Let’s break it down. Below you will find what is inside, how to get started, how the CME part works, who can and cannot use it, and where I would stay skeptical. One note up front: this post is built from OpenEvidence’s announcements and independent coverage. I have not used Learn hands on yet.
OpenEvidence Learn at a Glance
Learn lives at openevidence.com/learn. OpenEvidence says it is free to all of its users and was developed with medical institutions, certifiers and educators. It draws on the same peer reviewed evidence and clinical guidelines that power the company’s answers.
- Advantage: Study material starts from the questions you already ask in practice, so the topics are relevant by default.
- Limitation: The launch material shows what exists, not whether it improves exam scores or patient care. No outcome data has been published.
- Ideal users: Verified US trainees and clinicians who already use OpenEvidence and want practice questions, flashcards and CME in one place.
- Use case: Board prep during residency, or turning a tricky clinic question into a lasting memory.
- Use case: Collecting CME credit as a side effect of work you already do.
What Is Inside Learn
| Tool | What it does | Good for |
|---|---|---|
| Test Prep | More than 50,000 board style questions covering medical school, residency, fellowship and specialty exams, including USMLE and COMLEX | Exam practice |
| Preceptor | Socratic, case based tutoring where you work up a simulated patient at your training level | Clinical reasoning |
| Flashcards | Cards generated from your OpenEvidence questions, missed Test Prep questions or any topic, reviewed at spaced intervals | Retention |
| CME | Credit for the clinical questions you already ask, with certificates kept in one place | Licensing and certification |
Two launch partners are named: Johns Hopkins University School of Medicine and the Perelman School of Medicine at the University of Pennsylvania. Their trainees and attending physicians will try the platform and send feedback. OpenEvidence says more tools for active learning and clinical reasoning will follow, built with medical schools and residency programs.
Vishal Patel, MD, the company’s VP of Clinical Partnerships and AI Education, framed the aim as encouraging continual learning and sound clinical judgment. That second part is the hard one, and I will come back to it.
Why This Launch Matters
Most study tools sit apart from real work. You finish a shift, open a different app, and drill questions someone else chose. Learn flips that. Your own clinical questions become the syllabus. A review from iatroX calls this connected workflow the strongest idea in the launch. Keep in mind iatroX is a competing product and says so in its own disclosure.
The audience is already there. OpenEvidence closed a $250 million Series D at a $12 billion valuation in January 2026, and later reporting put a newer round at $15 billion. The company says most US physicians use it daily and that independent studies show more physicians use it than all other AI platforms combined. Those are company claims, so treat them as marketing until you see the studies.
Even so, my read is simple. A question bank company has to win your attention first. OpenEvidence already has it. That is a real head start, and it explains why established names like AMBOSS are moving the other way, adding clinical answers to their study platform. The two are converging from opposite directions.
One more thing worth knowing. Free access is supported by advertising. iatroX notes the ads are labeled and separated from generated information. If you want a refresher on how ad funded AI works, my post on ChatGPT ad targeting and the settings to change covers the basics.
How to Get Started (Step by Step)
I could not open the Learn screens myself, so these steps stick to what OpenEvidence and independent coverage document. Menu names may shift.
- Go to openevidence.com/learn. The page asks you to log in or register, with email, Google, Microsoft and Apple sign in options.
- Verify your professional status. OpenEvidence’s app listing says users must verify as healthcare professionals, and the terms point to an NPI or another applicable identifier. Have yours ready.
- Start with Test Prep for your exam. Pick the exam you are actually sitting, such as USMLE, COMLEX or a specialty board, then run a short block to see how the questions feel.
- Open Preceptor for one case. Choose a specialty and setting, then work up the simulated patient. Commit to each answer before you ask for help.
- Build flashcards from your misses. The release says cards can come from missed Test Prep questions, your OpenEvidence questions, or any topic.
- Check the CME section. Look at which activities you qualify for and where your certificates are stored.
A Weekly Study Loop You Can Steal
This is my suggestion, not OpenEvidence’s, and I have not tested it. It fits how the tools are described.
- During the week, jot down every clinical question that made you pause. Even three or four is enough.
- Once a week, turn those into flashcards, then run Preceptor on one related case.
- Follow with a 20 question Test Prep block on the same topic. If you miss a question, make a card from it.
- Review cards for ten minutes a day. Spaced intervals do the rest.
- Write one sentence on what changed your thinking before you claim any CME credit. More on that below.
The same idea works outside medicine. If you are curious how another company approaches adaptive study loops, see my post on Google Study Notebooks in Gemini.
The CME Part: Read the Fine Print
The Learn release promises credit for questions you already ask, but it does not name an accrediting body or give credit hours. For that, look at the July 28, 2026 announcement of OpenEvidence’s education platform. It says AKH, Inc. is the accredited provider, jointly accredited by the ACCME, ACPE and ANCC. It covers NPI verified physicians, nurse practitioners and physician associates. Maintenance of Certification credit is for physicians only, through select boards the release does not name.
Credit is not automatic. According to that release, you earn it by engaging with evidence based content, considering its relevance to your practice, identifying a key learning point, and documenting how it informs your reasoning or future care decisions. You then claim credit from a personalized dashboard. Physicians can track CME on the ACCME’s CME Passport, and you can enable automatic reporting or download a transcript to self report.
My advice: assume the Learn CME runs on the same system, then confirm it inside the product. Before you count on any credit, check that your state, employer or certifying board accepts AKH activities. That check takes two minutes and saves an unpleasant surprise at renewal time.
Who Can Actually Use It
This is a US first product. iatroX reports that OpenEvidence’s updated terms exclude access from the UK, EU and EEA, and registration relies on an NPI or another applicable professional identifier. A separate September partnership supports free access for verified Canadian physicians. Even where you can log in, an account is not the same as credit that your local regulator accepts.
The question bank is also built around US exams. If you are preparing for a different licensing path, the content may not line up. iatroX specifically says the launch does not establish alignment with the Canadian MCCQE curriculum. Check the syllabus before you spend weeks on it.
The Honest Limitations
- No outcome evidence yet. Nothing published shows better exam results, higher pass rates or improved clinical performance. Hopkins and Penn are giving feedback, which is not the same as a learning study.
- Question quality is unproven. Good questions force a real decision, and good explanations show why the wrong answers tempt people. Test a few blocks before you trust it as your main bank.
- An AI tutor can do your thinking for you. A tutorial can feel productive while the system supplies most of the reasoning. Answer first, then check.
- Auto generated flashcards can lose the details that matter. Severity, comorbidity and local guidance often decide the right management. Read each card for those qualifiers.
- A question is not a complete CME activity. Reflection is part of earning credit, so the certificate should reflect real learning, not replace it.
- Student access is unclear. The sources I found say users must verify as healthcare professionals. How trainees without an NPI sign up is not spelled out in what I read.
- Most claims come from the company. Usage figures and the 50,000 question count are OpenEvidence’s own statements.
Alternatives Worth Knowing
- UWorld: A paid question bank widely used for USMLE prep. It has a long track record, though it sits apart from your daily clinical questions.
- AMBOSS: Question banks linked to a medical library, now with an AI mode for clinical care. Stronger on structured study.
- Anki: Free, open source spaced repetition flashcards. You build the cards yourself, which takes effort but forces you to think.
The difference in one line: UWorld and AMBOSS start from the exam, Anki starts from your own cards, and Learn starts from the question you asked at work.
The Bottom Line
Learn is a smart idea with a built in audience. Turning real clinical questions into practice, flashcards and credit is a better loop than studying in a separate app. What it has not shown yet is that the loop works. If you are eligible, try one week with the study loop above and keep a paid question bank until the results justify dropping it.
Know a resident or med student drowning in board prep? Send them this post before their next study block.
Sources
- Business Wire, Announcing Learn: OpenEvidence’s Personalized Medical Education Platform.
- OpenEvidence, Announcements.
- Business Wire, OpenEvidence Launches New Education Platform Offering CE and MOC Credit.
- Unite.AI, OpenEvidence Launches Learn, a Free AI Medical Education Platform.
- iatroX, OpenEvidence Learn Review: Test Prep, Preceptor, Flashcards and CME.
- iatroX, OpenEvidence, AMBOSS and the Future of Medical Learning.
- eWeek, OpenEvidence Secures $12B Valuation.
- AI Weekly, OpenEvidence Raises $250M at $15B.

Leave a Reply