32B

GLM-5.5 Training Pipeline: 32B Model Could Lower Deployment Barriers

Zhipu AI Chief Scientist Tang Jie detailed GLM-5.5’s training pipeline, spanning 32B, 122B, and 256B parameters.