DIAGENS Sets New Medical AI Standards with Enhanced CT Analysis and iMed-6.0 Innovations

DIAGENS Revolutionizes Medical Imaging with iMed-6.0

On November 12, 2026, DIAGENS (HKEX: 2526) made headlines by announcing their latest benchmark achievements for iMed-6.0, a pioneering advancement in the realm of medical imaging. This state-of-the-art medical imaging foundation model is designed for comprehensive analysis, significantly enhancing the capabilities of CT imaging.

A Breakthrough in CT Analysis

DIAGENS introduced a new CT model powered by iMed-6.0, capable of analyzing a remarkable 509 diseases and imaging findings. This breakthrough allows for detailed assessments of both chest and abdominal CT scans, featuring a structure that identifies 263 maladies related to chest imaging and 391 related to abdominal examinations. This groundbreaking development marks a significant step forward in medical imaging technology.

During the testing phase of MedBench's multimodal API evaluation, both the iMed-6.0 models, Thoth and Aegis, excelled, achieving first and second place in multiple core categories. Thoth notably surpassed previous benchmarks on the Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark (MMMU) test, showcasing its exceptional capabilities.

Dr. Ning Song, the Chairman and CEO of DIAGENS, expressed the company's vision, stating, "Our aim is to empower clinicians and hospitals to build AI solutions tailored to their specific challenges. With our medical imaging foundation model and no-code tools, they can develop customized models that integrate seamlessly into their workflow and support clinical decisions."

Leading the Way in Medical AI Evaluation

MedBench, created by the Shanghai AI Laboratory, allows for rigorous evaluations within the OpenCompass evaluation ecosystem. During the assessment, iMed-6.0's models achieved impressive scores; Thoth garnered an overall score of 64.4, while Aegis attained 61.7. In talent areas such as medical visual perception and text extraction, these models exceeded performance expectations, scoring 88.8 and 84.9, respectively. These achievements underscore DIAGENS’s commitment to setting new standards in medical AI.

In another noteworthy accomplishment, Thoth achieved an excellent overall accuracy of 66.2% on the MMMU test, outpacing the previous best result of 65.9%. Within the Health and Medicine category, Thoth and Aegis achieved remarkable scores of 74.9% and 71.1%, respectively, establishing themselves as leaders in AI-driven analysis.

Enhanced Functionality for Clinicians

The newly developed CT model provides unique functionalities tailored to structured imaging settings, specifically non-contrast and contrast-enhanced CT examinations. A significant aspect is its automated identification capacity, categorizing a plethora of diseases and imaging findings.

DIAGENS conducted rigorous internal validations using nearly 40,000 analysis-report pairs from various hospitals, achieving a macro-averaged Area Under the Curve (AUC) of 0.855 for zero-shot classification of all 509 targets, and an impressive 0.930 on a score of 249 key targets represented in established benchmarks.

Building a Seamless Workflow for AI Development

The iMed-6.0 model excels at processing diverse imaging data, including 3D imaging data and whole-slide pathology images. This versatility allows for extensive interpretation, anatomical analysis, and report generation, significantly improving clinicians’ workflow.

In tandem with iMedLoop™, DIAGENS’s dedicated AI development platform, the organization provides comprehensive support for data preparation, model training, evaluations, and feedback options. This encourages hospitals and clinicians to build tailored AI models that best suit their specific imaging tasks.

As of mid-2026, DIAGENS has engaged in 158 model projects and forged RD partnerships with 99 medical institutions, reaching over 400 healthcare establishments. The company's innovative approach not only enhances standard practices but also democratizes access to advanced medical AI technologies.

Conclusion

DIAGENS continues to push the boundaries in the field of medical imaging AI, delivering tools that empower healthcare professionals and improve patient outcomes. The launch of iMed-6.0 marks a pivotal moment in the evolution of medical imaging analytics, showcasing the potential of AI to facilitate groundbreaking advancements in healthcare.