Neuro-VL

Neuroimaging vision-language model (Neuro-VL-8B-Thinking, Qwen3-8B language model with an HLIP visual encoder) that reads complete MRI and CT exams, converses about findings, drafts reports with reasoning traces and compares serial exams; trained on nearly 1 million clinical neuroimaging studies from 450,000 patients; released with the NeuroBench benchmark

LICCC BY-NC-SA 4.0UPD2026-11-13STATUSMaintainedCOMMNon-commercial
What it is
Vision-language model (Foundation models & agents)
Modality
Neuroradiology: brain MRI and head CT
Maintainer
University of Michigan MLiNS Lab (Zhao, Hollon et al.)
License
CC BY-NC-SA 4.0
Commercial use
Non-commercial
Access
Free, public source repository (see links below)
Last update
2026-11-13 (last repository push, UTC; Maintained)
Last checked
2026-11-16
Limitations
Listed for research and development reference. Hounsfield Lab has not tested its clinical performance; check the project’s own documentation for intended use and regulatory status.

Links

License review: Repo LICENSE and the Hugging Face model card (mlinslab/neuro-vl-8b-thinking, about 34 GB) are CC BY-NC-SA 4.0, non-commercial research only. The weights are gated: you must log in and accept the conditions, and requests are approved manually. The NeuroBench evaluation code (MLNeurosurg/neurobench) is Apache-2.0 and the NeuroBench dataset is CC BY-NC-SA 4.0. Paper: arXiv 2611.15094 (Nov 11, 2026).

Commercial use: read the project’s license (and any model-weight terms) yourself before relying on it. This isn’t legal advice.

Related tools in Foundation models & agents

Related public imaging datasets

Browse all DICOM datasets with license and access details.

← Back to directory