Google makes Gemini 4 Argon globally available

Google says Gemini 4 Argon is globally available through its API, AI Studio, Gemini platform, Antigravity, the Gemini app and AI Mode in Search. The public model launches at an introductory $2 per million input tokens and $10 per million output tokens, with a stated 1M unified input-output token limit. Why it matters: This is a routing and cost-planning change, not independent evidence that Argon will outperform a team’s current model. The introductory price ends November 30, so compare one fixed workload against the model you already use, including latency, tool reliability and the post-introductory rate.

Try this: Run one representative coding or research workflow on Argon and your current model with the same acceptance test; log input/output tokens, latency, tool failures and the $4/$20 post-introductory rate before changing routing.