Imaging AI software developer HOPPR has integrated two NVIDIA open models into its HOPPR AI Foundry platform for building, refining, and validating medical imaging AI applications, the company announced at NVIDIA GTC 2026.
The first model, NV-Reason, introduces multimodal reasoning capabilities for chest x-ray interpretation workflows, generating structured analytical reasoning steps alongside model outputs to provide greater transparency into how imaging interpretations are produced, according to the company. The second, NV-Generate, is a latent diffusion model that creates high-fidelity synthetic DICOM imaging datasets that support data augmentation and model training where real-world data may be limited, HOPPR said.
The HOPPR AI Foundry is built on NVIDIA A100 and H100 GPUs and supports optimization with NVIDIA TensorRT and inference through NVIDIA Triton Inference Server, the company noted.
