On October 13, Google released Gemini 4 Argon globally for paying API customers and businesses. The model was announced on September 30, but only a small group of cybersecurity specialists could try it at first.
Argon is designed for long, multi-step tasks such as large code changes, legal research and scientific problem-solving. Google also says it can produce much longer outputs than earlier Gemini models. That does not automatically make it better for everyday questions; a smaller model may be both faster and cheaper for those.
Access matters more than leaderboards here. Developers can now use Argon through the paid Gemini API. Google's pricing page explicitly says the model is not available on the free tier or Google AI Pro. Introductory pricing is $2 per million input tokens and $10 per million output tokens through November 30. The prices double after that.
So a free account is not enough if you want to chat with it. Google's phased launch and its own tests also do not show how well the model will work in your work. Independent comparisons will become more useful now that more people can actually try it.