Open-source medical video understanding system uAI-NEXUS-MedVLM released
让大模型理解真实医疗视频,全球首个开源技术方案来了!
United Imaging Intelligence released uAI-NEXUS-MedVLM for medical video understanding, with a CVPR 2026 paper. MedVidBench has 532k video-instruction pairs across 8 medical sources and 8 tasks. Qwen2.5-VL-7B SFT reached 89.4% CVS accuracy; GPT-5.4 scored 16.4%.
Why it matters: United Imaging Intelligence open-sourced uAI-NEXUS-MedVLM, a large model for ultrasound, endoscopy and other medical video, plus MedVidBench, an evaluation set of 532,000 video instruction pairs.