OpenEvidence, a startup that funds its AI tools with advertising revenue, is now valued at about $15 billion. The company’s software is being used by doctors who choose it on their own, according to investors. OpenEvidence markets the system as a literature‑based drug‑discovery platform that mines scientific papers for insights.
The startup has attracted notable talent, including Alexander Tarashansky, creator of Anthropic’s Claude Science, to help with clinical work. Investors such as Thrive Capital and the Mayo Clinic are backing the company, hoping its differentiation will fend off competition from big players like OpenAI and Anthropic.
How does it perform in real clinics?
A Mayo Clinic pilot that looked at five patient cases reported that OpenEvidence gave accurate, evidence‑based recommendations each time, though the impact on actual treatment choices was modest. By contrast, a June 2026 paper in *Nature* found the system lagged behind general‑purpose models such as Google’s Gemini and OpenAI’s GPT on standard clinical benchmarks.
Doctors such as NYU Langone neurosurgeon Eric Oermann note a lack of detailed research on individual AI tools and limited guidance for integrating them into everyday practice. There are also concerns that junior physicians might rely too heavily on the software without mastering basic medical knowledge.
Why it matters
If OpenEvidence’s recommendations are reliable, doctors who already use the tool could get faster, literature‑backed guidance without extra cost. However, mixed study results mean patients may not see a clear benefit yet, and clinicians need to verify AI advice before acting.
