.svg)
.svg)
.svg)
LiveDOSE Benchmark v1
Our first published benchmark. An open, reproducible evaluation with the full method, dataset and per-model results.
Check it outResearch
How we build, test, and understand
AI for life sciences.
From the questions
to the evidence.
Benchmarks · Studies · Engineering
Benchmarks
Live“We spend less time finding answers. And more time asking better questions.”
+15.1 pt lead
Select a model for details
We study the parts of pharma AI where a small error has a large consequence
Clinical-grade voice
Pronunciation, medical terminology and adverse-event capture in real-time speech.
Agents in regulated workflows
How agents stay inside approved content, and how to measure when they don't.
Context and identity
Resolving one physician across CRM, claims, medical and congress data.
Simulation and digital twins
When a simulated HCP panel predicts the real one, and when it doesn't.
Everything we've published
Open evaluations we publish and keep running.
Aug 2026DOSE Benchmark: Drug-name Oral Synthesis EvaluationNine TTS models, 274 drug names, 2,466 scored utterances. Synthio Labs leads at 91.2% pass rate.BenchmarkAug 2026DOSE recordings and scoring rubricEvery utterance, every score, and the pharmacist review notes behind them.DatasetSept 2025Evaluating a Medical-Grade Voice AI for Patient and Caregiver Guidance: A Multi-Scenario Nurse Panel StudyNearly 1,000 U.S.-licensed nurses role-played patients and caregivers in more than 20,000 interactions across 30 scenarios.Paper
No blog entries in this research index yet.Browse the blog →
