research
Scalable Benchmarking of Health AI’s Differential Diagnosis Accuracy
Our peer-reviewed framework evaluates August across 400 validated clinical vignettes spanning 14 medical specialties, reaching 81.8% top-one diagnostic accuracy and 95.8% specialist-referral accuracy
benchmark
August Scores 100% on the USMLE
August becomes the first health AI to score a perfect 100% on the USMLE, with leading results on MedQA and MMLU medical subsets.
benchmark
August's Perfect HealthBench Score
We aced HealthBench's emergency test. Now we're raising the bar.
benchmark
August AI Achieves 94.8% on the USMLE
August AI scores 94.8% on the USMLE — the highest of any benchmarked AI, beating GPT-4, MedPaLM 2 and OpenEvidence.
