The ARISE Healthcare Network – AI Research and Science Evaluation – was established by Stanford Medicine in 2024 as a collaborative effort among Stanford and other academic medical centers to advance healthcare AI with rigorous research studies.
ARISE recently announced its new framework, the Medical Artificial Intelligence Superintelligence Test, or MAST v1.0, which aims to measure and benchmark the path of AI models across clinical and general domains.
By launching the MAST framework simultaneously, their goal is to coordinate, maintain and rapidly update the highest-quality AI benchmarks within a shared infrastructure that enables cross-benchmark inferences to characterize model performance and underlying traits.
Together, MAST-Clinical domains, which cover diagnosis, management and safety, and MAST-General domains, which trace radiology and multimodal reasoning capabilities, enable the researchers and others to compare models against others within their classes.
The launch version of the MAST platform maintains evaluations and standardized model runs with public reporting across diagnosis, management and other domains, but future iterations will cross-benchmark trait analysis and use benchmarks as probes of underlying clinical behaviors.
【MORE】