Availability differs across Google's products
Google's October 13 announcement expands access to Gemini 4 Argon for developers, enterprises and consumers. The company describes API access through AI Studio and Gemini platform, Antigravity access for Google AI Ultra subscribers, and Gemini app and Search AI Mode access for Ultra and Pro subscribers.
Those product routes and account requirements should be considered separately. The announcement's broader-access wording does not mean every route is available without a subscription or usage charge.
Pricing and examples are evidence with limits
Google lists introductory API pricing at $2 per million input tokens and $10 per million output tokens, with a stated 95% discount for cached input tokens. An application's bill still depends on its requests, output volume and whether inputs qualify for caching.
The company presents engineering examples involving refactoring, performance optimization and visual code generation, along with professional and research workflows. It attributes improvements to work performed with Argon and publishes links to supporting changes and demonstrations.
These examples illustrate specific projects. They do not establish that every codebase, professional task or model comparison will reproduce the same improvement.
Compare the complete task rather than one answer
Franklin analysis: select a task with an existing baseline and a result that can be checked. For a code optimization, preserve the original behavior tests and measure performance after reviewing the proposed change. For a document workflow, retain the evidence supporting each important conclusion.
Record model access, tool use, total tokens and elapsed time together. This connects Google's access and pricing announcement to a usable evaluation of a whole workflow. A model that can attempt deeper work is valuable when the outcome remains correct and reviewable, with the account and resource requirements understood before routine automation is enabled.