Gemini 4 Argon reaches the app and API at an introductory $2-per-million input rate

What changed

Google’s October 13 launch makes Gemini 4 Argon available through its API to developers and enterprises, and through the Gemini app, Search AI Mode and Antigravity for eligible subscribers. In the app, a task toggle lets the model work across more than 40 connected services, including booking, shopping and design tools.

The developer API launches at an introductory $2 per million input tokens and $10 per million output tokens, with the company saying regular pricing begins after November 30. Google positions Argon for complex multi-step tasks in coding, finance, law, cybersecurity and science. Those are the company’s intended use cases, not a guarantee that each workflow will run reliably without supervision.

What the result means

The important shift is distribution: a frontier model is showing up in a consumer chat surface and connected-app tasks as well as developer tools. Access is tiered, and the launch includes a new model rather than universal free use. For readers, the useful question is whether task mode completes a bounded job with clear progress and controllable actions, not simply whether the model can reason at length.