Corti Launches Agentic AI Model for Medical Coding

Symphony for Medical Coding, an AI model claiming 25% higher accuracy than OpenAI, Google, and Anthropic systems in clinical coding benchmarks.

Corti has released an agentic artificial intelligence model for medical coding, stating that it outperforms competing systems from major technology providers by more than 25% in clinical accuracy benchmarks.

The model, titled Symphony for Medical Coding, is available via API and builds on the company’s existing Symphony platform, which is currently used by over 200 healthcare teams in the United States. Corti provides AI solutions to electronic health record vendors, virtual care platforms, practice management systems, and life sciences organizations globally.

Medical coding, particularly under the ICD-10 system with approximately 70,000 diagnosis codes, requires converting unstructured clinical notes into standardized data used for billing, research, and policy. Increasing adoption of ambient AI scribes has led to longer and more detailed clinical documentation, raising the risk of missed or inaccurate coding.

According to Corti, its model was evaluated on established benchmarks, including ACI-BENCH and MDACE, using identical conditions across all competing systems. The company stated that its model outperformed systems from OpenAI, Anthropic, Google, Amazon, and Oracle, including models such as Claude Opus 4.6, GPT-5.4, and Gemini 3.1 Pro, as well as derivative enterprise solutions.

Corti’s approach differs from traditional coding automation systems that rely on pattern recognition from annotated datasets. The company said its model uses a multi-agent architecture designed to replicate the reasoning process of human coders. The system includes agents that identify clinical evidence, interpret coding hierarchies, validate against guidelines, and resolve ambiguities.

In a study using Danish patient data, Corti identified three times as many suicide attempts as were previously coded, highlighting potential gaps in traditional coding systems and their impact on clinical insights and resource allocation.

The model has also been validated on real-world clinical data from a large U.S. health system across emergency and outpatient settings. Corti stated that all models in the comparison were tested multiple times to ensure consistency and reproducibility.

The platform is compliant with HIPAA and GDPR standards, with increasing demand from U.S. customers for higher data privacy safeguards. Founded in 2016, Corti has raised $100 million to date and continues to focus on reducing administrative burden and improving clinical workflows through AI-driven automation.

Stay tuned for more such updates on Digital Health News