gemini-cli accepts invalid model names with --model flag

[JaisalJain]

What happened?

When running the CLI with an invalid model name using the --model flag, the CLI starts normally instead of failing early.

Example:

gemini --model gemini-4-pro

gemini --model gemini-5-pro

Even though the model does not exist, the CLI launches successfully. The error only appears later when a request is made.

This can confuse users because it is not clear that the model name is invalid during startup.

What did you expect to happen?

The CLI should validate the model name during startup and exit with an error if the model does not exist.

Example expected behavior:

gemini --model gemini-4-pro

Invalid model: "gemini-4-pro"

Client information

Details

CLI Version - 0.32.1
OS - linux (ubuntu24.04)
login Using Google Account login.

Login information

No response

Anything else we need to know?

No response

[JaisalJain]