Meta Opens Muse Models to Caltech, Ohio State and Other Universities
Meta is giving researchers at Caltech, Ohio State's College of Medicine and three other US universities its latest Muse models, with $2 million in grants.
Until now, Meta's Muse has mostly shown up as a personal AI assistant. This time it is heading into university labs. Meta has announced partnerships with five US universities, giving researchers its latest Muse models along with funding and collaboration with Meta's own scientists.
Five schools, five different uses
The five are Caltech, The Ohio State University College of Medicine, Louisiana State University, Indiana University and UTHealth Houston. Meta says it is granting $2 million across the five schools in total, not per school. Support varies by institution and project and can include model access, work with Meta researchers and funding for select projects.
The models on offer include Muse Spark Open, which Meta calls its most capable open-weight model, the Muse Wonder and Muse Spark series, and Muse Glimmer, a 30B open-weight model that runs on a single consumer GPU. Vision models such as SAM, V-JEPA and DINO are part of the mix too.
Caltech's focus is fundamental science. Researchers get frontier Muse models through Meta's first-party API to test whether large language models can help formulate and prove conjectures and predict unseen outcomes in mathematics and the physical sciences, not just summarize existing knowledge. Caltech says five faculty members receive research grants for up to three years.
The medical projects are concrete. LSU is building a vision system to detect displaced tracheostomy tubes in infants, Indiana University is working on conversational AI for people with dementia, and Ohio State's College of Medicine is trying to predict heart failure risk from routine abdominal CT scans and other data.
Clear limits on data
One point matters for the healthcare work. Meta says universities will run its models inside their own environments, conduct the research independently and keep control of patient and clinical data, which will not be shared with or processed by Meta.
These are research plans and early prototypes, still far from clinical use. For Meta, it is a smart move to carry Muse's reputation from consumers into science. For the schools, it means frontier models without handing over data. The next thing to watch is which project produces reproducible results first.