Symphony for Medical Coding delivers production-ready medical coding automation across the US and Europe, based on learnings from the largest study of its kind.
NEW YORK and COPENHAGEN, Denmark, April 1, 2026 /PRNewswire/ — Corti, the frontier lab for clinical-grade AI, today released Symphony for Medical Coding, an agentic model that outperforms OpenAI and Anthropic – as well as Amazon, Oracle, and Google – in medical coding by more than 25% in clinical accuracy benchmarks. It is now available via Corti’s API to any team building AI-powered healthcare software.
The cost of getting it wrong
Medical coding converts clinical reality into structured data, powering reimbursement, reporting, and public health decisions. Coding errors are expensive, but the human cost goes much further.
One example shows the scale of what is missed: in a recent study of Danish patient data, Corti identified three times as many suicide attempts as had been coded. The cases were all there – recorded in clinical notes, flagged in medication records – but coders, working under time pressure, had missed them. When cases go uncounted, health systems can’t monitor trends, allocate resources, or design interventions. Policy fails before it starts.
Defined by frontier research
Medical coding is fundamentally a reasoning task, not a prediction problem. It involves interpreting many complexities, real judgment, and justification across thousands of codes. The American coding system alone, ICD-10-CM, has 70,000 diagnosis codes. Even worse, coding is based on guidelines that constantly evolve, making historically data-trained models inadequate.
Corti started addressing this by conducting the largest study of its kind (5.8 million patient encounters), leading to Code Like Humans, a multi-agent framework accepted to EMNLP 2025, one of machine learning’s top conferences. This framework mirrors professional coders’ steps: identifying evidence, reasoning through hierarchies, validating against guidelines, and reconciling ambiguity. Symphony for Medical Coding builds on this foundation to perform work like expert coders, delivering higher quality than other models at a fraction of the cost.
“Most AI systems fall short in medical coding because they treat it as labeling, not reasoning. Correct coding depends on evidence, context, hierarchy, and guideline interpretation. We built Symphony for Medical Coding to follow the same decision process expert coders use, and that is why the performance gap is so meaningful,” said Lars Maaløe, PhD, CTO and co-founder of Corti.
