Model aliases

Kinference allows you to provide aliases for long-form model checkpoint names, such as “glm” for “zai-org/GLM-5.3” or “qwen-3.6” for “Qwen/Qwen3.6-35B-A3B”. Inference consumers can then request models by either full checkpoint name or alias, allowing the platform to provide specific quantizations to users who require them, and abstractions over checkpoint detils for users who simply need the broader model family.

Model aliases are configured in the InferenceWorkerSet resource by setting the aliases field in the models section of the specification. See Basic Inference Serving for an example walkthrough.