Skip to main content
Serverless quickstarts need a model name such as MiniMaxAI/MiniMax-M2.5.

​
How To Discover A Serverless Model

If you have Control Plane access for the same Tensormesh environment, log in first and then use the pricing command: 
tm auth login
tm billing pricing serverless list
For machine-friendly output: 
tm --output json billing pricing serverless list --active-only
Use the returned pricing[].model value in SDK, CLI, or raw API requests.
  • Control Plane pricing records can help you discover published serverless models for the current Tensormesh environment.
  • Serverless requests use a model name string, not a Control Plane modelId UUID.
  • The pricing[].model field is the value to pass as model.
  • The id field is the pricing record id, not the inference model name.
  • The name field is display-oriented; use model when sending requests.

​
What To Do In Practice

  • If your team already has a known serverless target, use that exact model value in the SDK, CLI, or raw API request.
  • If you have Control Plane access for the same environment, run tm billing pricing serverless list and pick the model from the returned records.
  • If you only have inference credentials, or you are targeting a different serverless host or base URL override, ask your operator or admin for the exact serverless model string for that host before sending the request.

​
Alternative Path

If you decide you would rather use the standard On-Demand setup flow instead of serverless discovery, switch to tm auth login plus tm init --sync and use the synced served gateway model name for that deployment.