llm-gemini adds Argon to command-line and Python workflows

llm-gemini 0.35 adds Gemini 4 Argon to the LLM tool’s existing model interface. Developers can select it from the command line or Python and choose low, medium or high reasoning effort. This is an integration update to an already announced model. Using it requires Gemini API access and a configured key; the plugin does not supply model access itself.

Simon Willison · Oct 13, 2026

Why it matters

An existing scripting interface makes comparing a new model easier than building another client.

More context

Install the plugin in the same environment as LLM and configure a Gemini key. Try representative tasks at different reasoning levels and compare output quality and usage, rather than relying on illustration examples.

Keep in mind

Model availability and usage charges remain with the provider, independently of the plugin.

Original read