Google starts rolling out Gemini 4 Argon, but only for paying users

In short: Google has started rolling out its Gemini 4 Argon AI model, but access is mostly limited to paid plans and higher tiers get more features.

What happened

Google is beginning a wider rollout of Gemini 4 Argon starting today. Argon is Google’s latest AI model, meaning the engine that powers responses in tools like the Gemini app.

For everyday questions in the Gemini app, Argon will be available to people on Google’s AI Pro plan, which costs $20 per month, and the AI Ultra plan, which starts at $99 per month. Free users are not included. Google has recently limited free access to some other Gemini models as well.

Access is tighter in Google’s coding-focused products. In Antigravity and AI Studio, Argon is only available to AI Ultra subscribers. Pro subscribers can still use Argon to generate code inside the Gemini app’s Canvas mode, but not in Antigravity’s more complex “agent” workflows, which are like asking the AI to carry out a longer set of steps on your behalf.

Google also shared internal examples of Argon’s use. It says Argon helped make YouTube’s comments system more efficient by finding duplicate emoji work, reducing server load by 45 percent. Google also says Argon found long-standing bugs in a drone simulator used by Project Wing.

Why it matters

This rollout shows that more capable AI features are increasingly being put behind higher-priced subscriptions, especially for coding tools that can be expensive to run. For developers using Google’s paid API (a way for apps to connect to Gemini), Google is offering temporary discounted pricing through November, before higher rates kick in later.

Source: Arstechnica