Google has released its new flagship AI model, Gemini 4 Argon, featuring improved reasoning capabilities and competitive pricing for subscribers and developers.

Google has launched Gemini 4 Argon, its latest frontier AI model, which is now available globally to Pro and Ultra subscribers and via API. The company claims the model offers superior reasoning and performance across complex tasks, including coding and cybersecurity. Argon replaces the Gemini Pro class as Google's new flagship series, with the company positioning it as a direct competitor to other industry-leading models.

Availability and Access

Gemini 4 Argon is accessible through the Gemini app, AI mode in Search, and the Antigravity development platform for Ultra and Pro subscribers. Developers can also access the model via the API in AI Studio. While the model is now the company's primary frontier offering, free users and Plus subscribers currently retain access only to Flash-tier models.

Performance and Internal Use

Google reports that Argon has already been deployed internally to assist with cybersecurity patching and code optimization. In the Chromium project, the model assisted in an architectural refactor of a 40-million-line codebase in weeks rather than months. Additionally, Google claims that software engineers using the model increased agent-driven code submissions by over 35 percent, while code rollbacks decreased by nearly a third.

Pricing Structure

The model is launching with an introductory price of $2 per million input tokens and $10 per million output tokens, with a 95 percent discount on cached inputs. These rates are scheduled to double at the end of November. Despite the planned increase, Google maintains that the pricing remains significantly lower than comparable frontier models from competitors.