Models - TypeScript SDK
client.models reads the deployable model catalog. Caller-owned assets are
managed through client.myModels.
Pass the catalog's modelId to get() and deploymentCapabilities(). Pass
its modelName unchanged when creating a Deployment; do not construct either
value from the provider label.
Overview
Available Operations
| Method | Description |
|---|---|
list(options?) | Search and page the model catalog. |
get(modelId) | Read one catalog model. |
deploymentCapabilities(modelId, options?) | Resolve supported deployment choices, optionally for a cluster. |
list
Search and page the model catalog.
Request
Optional keyword, modelType, provider, source, page, pageSize.
Response
PageResult<ModelInfoVO>.
get
Read one catalog model.
Request
modelId: string required.
Response
ModelInfoVO.
deploymentCapabilities
Resolve supported deployment choices, optionally for a cluster.
Request
modelId required; optional {clusterId}.
Response
ModelDeploymentCapabilitiesVO.
Field Reference and Examples
Key response fields
| Type | Fields |
|---|---|
ModelInfoVO | modelId, modelName, displayName, modelType, provider, source, modelSize, deployable, supportedGpuTypes, supportedServingModes, deploymentOptions |
ModelDeploymentCapabilitiesVO | modelId, modelName, clusterId, gpuTypeOptions, recommendedGpuType, acceleratorPools, clusterGpuTypes |
const page = await client.models.list({
keyword: "Qwen",
modelType: "LLM",
provider: "Qwen",
source: "gallery",
page: 1,
pageSize: 20,
});
const model = await client.models.get("qwen3-4b");
const choices = await client.models.deploymentCapabilities(
"qwen3-4b",
{ clusterId: "1" },
);
All methods use the shared authentication and typed error behavior described in Response Conventions and Retry and Security.