zai-org
GLM-4.5-Air
text generationmit1.3M downloadsHugging Face
No server in stock right now has the 247 GB this model needs for inference.
GPU memory it needs (estimate)
Rule of thumb from the model's size. Real needs depend on batch size, context length and framework.
Inference (full precision)
247 GB
Inference (8-bit)
123 GB
Inference (4-bit)
62 GB
LoRA fine-tuning
267 GB
QLoRA (8-bit base)
134 GB
Full fine-tuning
988 GB