OpenEvidence
AI clinical decision support for physicians — cited, evidence-based answers from peer-reviewed medical literature in seconds.
Overview
OpenEvidence is a free, ad-supported clinical decision-support platform for verified US physicians, built by ex-Kensho founder Daniel Nadler. It answers clinical questions with citations from NEJM, JAMA, and 300+ journals, reaching over 757,000 verified US physicians by mid-2026.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 4 Sep 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- NPI verification required
- Unlimited clinical Q&A with citations
- Ad-supported model
- Pricing undisclosed — contact sales
- System-level customization
- Major health systems currently on the free model
Use cases
What you can produce with OpenEvidence
- Cited, peer-reviewed answer to a clinical question with source links to NEJM, JAMA, or 300+ partner journals
- Deep literature synthesis report (DeepConsult) analyzing multiple studies for a complex clinical scenario
- Evidence-grounded prescribing guidance backed by current clinical guidelines from AMA, NCCN, ACC, or ACEP
- Voice-transcribed clinical Q&A session (hands-free mode) with written transcript and references for post-encounter review
- AMA PRA Category 1 CME credit certificate earned through in-platform educational content
- Direct in-workflow evidence lookup surfaced inside Epic EHR (for participating health systems)
ASEAN Perspective
OpenEvidence in Southeast Asia
OpenEvidence is effectively US-only in practice: access requires a US National Provider Identifier (NPI), which ASEAN and broader APAC clinicians do not hold, blocking the automated verification pathway. The company's published expansion roadmap targets English-speaking markets first (UK, Canada, Australia) with no stated timeline for Southeast Asian or broader APAC rollout, and no localization for regional clinical guidelines (MOH Singapore, Thai FDA, BPOM Indonesia, etc.). Data residency is US-centric with no disclosed PDPA, HIPAA-equivalent, or regional data-sovereignty arrangements for APAC users. Until OpenEvidence opens a non-NPI verification path and integrates ASEAN guidelines, clinicians in Singapore, Malaysia, Indonesia, the Philippines, Thailand, and Vietnam will find it of limited practical utility — though the free access model and quality standard it sets is closely watched by regional health-tech players and investors.
OpenEvidence is genuinely useful for US clinicians: free, citation-backed answers pulled from NEJM, JAMA, and AMA guidelines, funded by pharma advertising instead of subscriptions. Growth has been extraordinary — over 690,000 verified physicians and 915,000 total licensed clinicians as of July 2026, alongside a $250M Series D that pushed its valuation to $12 billion in January 2026.
The caveats matter more than the growth numbers suggest. A December 2025 preprint found accuracy on complex subspecialty cases fell to 41% for DeepConsult and 34% for standard queries — well below its widely cited perfect USMLE score. A peer-reviewed PMC study found it mostly reinforces clinicians' existing plans rather than changing them. There's no public API, EHR integration remains early-stage, and the pharma-ad funding model raises conflict-of-interest questions nobody has independently audited yet.
What people say
Ask physicians on Sermo what worries them about OpenEvidence and 44% say accuracy and the risk of misinformation; another 19% point to the lack of physician oversight or explainability built into the answers. That skepticism sits oddly next to the adoption numbers: over 690,000 verified US physicians and 915,000 total licensed clinicians were using the platform as of July 2026, per the company's own NewYork-Presbyterian announcement, with more than a million questions fielded some days.
The Clinical AI Report's physician survey rates it 4.3/5 across 16 reviews and ranks it #2 in clinical decision support — strong marks for clarity (3.55/4) and relevance (3.75/4). But the same peer-reviewed PMC study found its decision-impact score a much weaker 1.95/4, meaning it usually confirms what a doctor already planned to do rather than changing the plan. A December 2025 preprint (not yet peer-reviewed) tested it against complex subspecialty scenarios and found accuracy dropped to 41% for DeepConsult mode and 34% for the standard Quick Consult — a long way from the perfect USMLE scores that made headlines.
Clinicians who want to trust it more keep asking for the same thing: peer-reviewed validation (37% in the Sermo survey said this would move them) and institutional endorsement from hospitals or medical societies (21%). The pharma-ad revenue model, the US-only NPI verification gate, and the absence of a public API round out the recurring structural complaints.
Summary of public user & expert reviews, compiled by RECATOOLS.
Notable facts
- OpenEvidence's AI model scored 100% on the US Medical Licensing Examination (USMLE) in 2025 — the first clinical AI to achieve this benchmark.
- CEO Daniel Nadler previously built Kensho, an AI company acquired by S&P Global for ~$550 million — he has now applied the same literature-synthesis playbook to medicine.
- The platform generates revenue through pharmaceutical advertising at CPMs of $70–$1,000+, orders of magnitude higher than social media CPMs, targeting prescribing physicians at the moment of clinical decision-making.
- By early 2026, a physician aboard a commercial flight used OpenEvidence mid-air to determine whether a passenger's immunosuppression warranted an emergency landing — an anecdote that circulated widely in medical communities.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including OpenEvidence's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with OpenEvidence unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to OpenEvidence directly →
Spotted something out of date? Suggest an update →
Alternatives to OpenEvidence
More in Research & Data
