GLM is a low-cost option for the high-volume steps in a job: classification, extraction, and first-pass drafting.
Weights are open, so a step running on the hosted endpoint can move to hardware you control without changing how the job is written.
- Capabilities
- low cost per token · tool use · open weights