DIAGENS CT Model Targets 509 Diseases and Imaging Findings
DIAGENS has introduced a CT model designed to identify 509 diseases and imaging findings in chest and abdominal scans, alongside benchmark results for its iMed-6.0 medical imaging foundation model. The company says two iMed-6.0 models took the top two positions in MedBench's multimodal API evaluation.
- iMed-6.0 Thoth and Aegis ranked first and second in DIAGENS's MedBench evaluation.
- Thoth achieved 66.2% overall accuracy on the MMMU Test.
- The CT model supports 509 chest and abdominal diseases and imaging findings.
- Internal validation covered nearly 40,000 CT examination and report pairs.
- DIAGENS reported 158 model projects and partnerships with 99 hospitals as of June 30, 2026.
Two iMed-6.0 models lead MedBench results
DIAGENS said iMed-6.0 Thoth and Aegis ranked first and second, respectively, both overall and across MedBench's three core categories. MedBench is developed by Shanghai AI Laboratory and forms part of the OpenCompass evaluation ecosystem.
Thoth received an overall score of 64.4 in the multimodal API evaluation, while Aegis scored 61.7. The models also ranked in the top two for medical visual perception and text extraction, cross-modal semantic understanding and reasoning, and clinical decision support and reasoning.
In medical visual perception and text extraction, Thoth scored 88.8 and Aegis scored 84.9. DIAGENS said these results were 20.3% and 15.0% higher, respectively, than the highest score of 73.8 among the other evaluated models.
Thoth exceeds the reported MMMU benchmark result
On the Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark, known as the MMMU Test, Thoth achieved 66.2% overall accuracy. The result exceeded the previously published best result of 65.9%, according to DIAGENS.
The benchmark covers six broad domains and 30 subjects. In its Health & Medicine category, Thoth achieved 74.9% accuracy and Aegis achieved 71.1%, placing the two models first and second in the benchmark evaluation.
DIAGENS describes iMed-6.0 as a multimodal system that processes 3D imaging data, whole-slide pathology images and temporal imaging data. It supports image interpretation, anatomy and lesion analysis, and report drafting and review. For supported tasks, outputs can be linked to specific slices or regions in source images.
CT model covers chest and abdominal examinations
The new CT model is designed for non-contrast and contrast-enhanced CT examinations of the chest and abdomen. It supports automated identification of 509 distinct diseases and imaging findings, including 263 chest targets and 391 abdominal targets. DIAGENS said overlapping targets are counted once.
In an internal validation using nearly 40,000 CT examination and report pairs from multiple hospitals, the model achieved a macro-averaged AUC of 0.855 for zero-shot classification across all 509 targets. For 249 key targets represented in internationally available public benchmarks, it achieved a macro-averaged AUC of 0.930.
Zero-code tools are aimed at specialty AI development
The CT model was developed using iMed-6.0 Aegis. DIAGENS combines the foundation model with iMedLoop, its medical imaging AI development platform, which connects data preparation, model training, evaluation, deployment and feedback.
Its iMedMaaS service provides a zero-code interface for clinicians and hospitals to train, deploy and scale custom models. As of June 30, 2026, DIAGENS said it had undertaken 158 model projects, established research and development partnerships with 99 hospitals, and reached more than 400 healthcare institutions through its products and services.
Dr. Ning Song, Chairman and CEO of DIAGENS, said the company wants clinicians and hospitals to build AI around the clinical challenges they know best. The announcement positions the foundation model and zero-code tools as a way to develop specialty models, adapt them to clinical workflows and use them as copilots for clinical decisions.


