Quick clip of GLM 5.2 running locally on an Nvidia DGX Station via a VS Code coding agent, with the creator judging the output speed fast enough despite the model not fully fitting into HBM.
Continue to AI University →