Skip to main content
We have added cost calculation and attribution for Gemini CLI usage via Truefoundry Gateway - however, the models now need to be registered in the gateway for Gemini CLI to work. This change concerns you only if you are using Gemini CLI via the gateway. All models used by the Gemini CLI must now be added to the Gemini account registered on Truefoundry that is used to access Gemini CLI via the TFY AI Gateway.
Earlier, this was optional and we would just pass the model that came in the request from Gemini CLI even if that was not registered in Truefoundry. However, after v0.118.2, the requests will fail if the models from Gemini are not registered in Truefoundry.

​
Important

The model display name must exactly match the model ID, because Gemini CLI uses the actual model ID and Truefoundry treats that as the model name.

​
Example

  • Model ID: gemini-2.5-pro
  • Model name: gemini-2.5-pro (must be exactly the same)
This change enables cost calculation for Gemini CLI. After this, you’ll be able to see the cost for each request and view it in metrics as well.