Gemini 4 Argon lands in Figma Make: what it changes

Gemini 4 Argon arrives in Figma Make

A new model joins the lineup. Figma has announced the gradual rollout of Gemini 4 Argon, presented as Google's most advanced model to date, now integrated directly into Figma Make.

In practice, using it is nothing complicated: open the model selector in the Figma Make chatbox, pick Gemini 4 Argon, and you are off. The model is available on all plans, no exceptions.

What this actually changes for your prototypes

As a reminder, Figma Make turns a text prompt into a functional interface, with code generated live from a simple description. A new model landing in the selector is never a minor thing: each model has its own way of interpreting a prompt, structuring code, and handling interface complexity.

With Gemini 4 Argon, the idea is simple: widen the options. You are no longer locked into a single generation engine. You can test, compare, and keep whichever one best fits your workflow or the type of project you are working on.

Why this announcement fits a bigger trend

This release did not come out of nowhere. It adds to a series of recent updates that show the direction Figma is heading: more agents, more automation, more AI-assisted generation, whether in Figma Make, in the Figma agent now available in general release, or in Figma Motion, which keeps gaining depth. The common thread is clear: Figma wants you spending less time on execution and more time on creative decisions.

If these developments interest you, we already dug into AI agents and models in our article on Figma Motion and shaders, another piece of this same movement.

{{webdesign}}

Mazette.co's take

A new model in the selector, sure, why not. But let us be honest: this is not the kind of announcement that overhauls the way we work overnight. It is one more useful building block in an ecosystem that has been piling up quite a few of these lately.

What genuinely interests us at Mazette.co is what this enables upstream of a Webflow project. Figma Make is often used to quickly explore an idea, to test a direction before moving into the final design phase. The stronger the available models, the faster and more relevant that exploration phase becomes. Saving time on prototyping means more time to polish the user experience, where it actually matters.

Our vigilance remains the same as with any AI code generator: a generated prototype is never an end in itself. It gives you a direction, a starting point for discussion with the client, but the work of clean structuring, accessibility, and consistency with a design system remains essential. No model, however "Argon" it may be, replaces the eye of a designer who understands the real constraints of a site in production.

Our advice: try it on your next exploratory prompts, compare it with the models already in place, and keep your critical thinking sharp about the result. That is exactly the approach we take in our Webflow and product design projects every day.