Work with advanced language models in a secure research environment. Use AI for writing, analysis, idea development, and teaching.
Hosted at SDU with clear data governance — meaning sensitive research data stays within the university infrastructure and is not exposed to third-party platforms or reused for model training.
Chat UI requires a GPU to run AI models. The required GPU size depends on the size of the model you want to load.
For example, llama3.2:3b takes up around 2.0 GB and can easily run on gpu-nvidia-b200-1-mig.1g, which has 32 GB of GPU memory. Larger models, such as models around 70B parameters, may require more GPU memory and will typically not fit on this machine type.
