Corti, a clinical-grade AI laboratory, has announced the release of Symphony for Medical Coding, an agentic model designed to automate medical coding across the United States and Europe. According to the company, the model demonstrates a clinical accuracy edge of up to 25% over competitors including OpenAI, Anthropic, Google, and Amazon. The system is built upon the “Code Like Humans” multi-agent framework, which was developed following a peer-reviewed study of 5.8 million patient encounters.
The technology moves beyond simple predictive labeling by utilizing a reasoning-based approach that identifies evidence, navigates complex hierarchies, and validates codes against evolving guidelines. This methodology aims to reduce the high human and financial costs associated with coding errors, such as the underreporting of critical health trends like suicide attempts. Symphony for Medical Coding provides auditable results by linking assigned codes directly to clinical evidence and highlighting ambiguities for review by compliance leaders.
The model is currently available via Corti’s API and supports various standards including US diagnosis and procedure coding (ICD-10-CM, ICD-10-PCS, CPT) as well as European environments. It is designed to function across different international markets without the need for local retraining. Headquartered in Copenhagen, Corti currently serves over 100 million patients annually, and its new coding infrastructure is now accessible through the Corti Console for integration into healthcare software.
