What the EU AI Act means for healthcare and medical AI: the four risk levels, the timeline, when clinical software counts as high-risk, how it links to the MDR, and how to approach compliance.
Laboratory medicine industrialised measurement and left interpretation artisanal. Why static comment libraries fall short, and what draft-and-sign could fix.
The Surviving Sepsis Campaign 2026 guideline: 129 statements, 46 new. What changed since 2021 on screening, antibiotic timing, MAP target, and vasopressors.
MTS-Bench compared ChatGPT (GPT-5.1), a general-purpose LLM, with DR. INFO against the Manchester Triage System. ChatGPT undertriaged 44.2% of cases; DR. INFO 11.5%.
Give a medical AI model a standard case vignette and it performs strongly. Turn the same cases into interactive clinical conversations and some models keep less than a tenth of their diagnostic accuracy. The test format matters as much as the number.
.png)
A benchmark score looks precise, yet a single percentage can hide what matters most to a clinician. A practical guide to reading clinical AI benchmarks — what they measure, how they are built, and twelve questions to ask before trusting the leaderboard.
Takeaways from the first English DR. INFO and ZAKA webinar on clinical AI. Dr. Mohamad Diab and Dr. Nader Absi on where AI helps, where it falls short, and why better prompting is really clinical reasoning, with three cases and a live Q&A.
The Portuguese Society of Internal Medicine (SPMI) and DR. INFO are launching a collaboration for the responsible use of clinical AI in Portugal, easing access to reliable, source-based medical information while promoting reflective use of AI in everyday clinical practice.
A practical summary of how AI can support doctors with telephone assistants, documentation, prescription workflows, and evidence-based medical information.
A prospective pilot study published in Cureus reports positive physician perceptions of DR. INFO as an agentic AI clinical assistant at the point of care, with high ratings for perceived time saving and decision support.
The bar for clinical AI is rising fast. Regulatory timelines are hardening, international health bodies are tightening ethical frameworks, and independent benchmarking is becoming the expected standard. Here is where DR. INFO stands.
Synduct GmbH announces that DR. INFO has been CE-marked under the European Union Medical Device Regulation (EU MDR), confirming essential safety and performance requirements for regulated deployment across European healthcare systems.
DR. INFO scored 0.68 on HealthBench Hard, outperforming GPT-5, GPT-5.2, Gemini 2.5 Pro, Grok 3, and every clinical AI system evaluated - built on real evidence, cited in real time, designed for real medicine.
We rebuilt the interface from the ground up - faster, more intuitive, and better aligned with how clinicians search, validate, and apply medical information in real practice.
DR. INFO was on-site at ASTOR 2026's 33rd Congresso da Medicina da Dor, with CMO Dr. Rogério Corga da Silva as an invited speaker in the AI session alongside ChatGPT, Google, and Microsoft.
Programmatic access to evidence-grounded medical search powered by Large Language models. Built for developers with privacy by design, streaming responses, and transparent citations to authoritative medical sources.
.png)
DR.INFO achieves a 0.51 score on HealthBench's Hard subset, demonstrating advanced clinical reasoning and safety-aligned conversational behavior across 1000 complex real-world scenarios.
DR.INFO achieves 95.4% accuracy on USMLE-style benchmark, exceeding expected performance for human medical trainees and demonstrating strong medical knowledge depth across major clinical domains.
Experience the leading medical AI platform trusted by healthcare professionals worldwide