Home / Technology / Google's Gemini 4 Argon: Power Unleashed
Google's Gemini 4 Argon: Power Unleashed
13 Oct
Summary
- Gemini 4 Argon AI model is now broadly available to developers.
- Google reports significant internal productivity gains with Argon.
- Argon shows competitive performance in professional tasks.
Google made its flagship Gemini 4 Argon multimodal AI model broadly available on October 13, 2026. This powerful new option for software engineering, financial analysis, and scientific research was previously restricted to early testers. Argon is now accessible via the paid Gemini API, Google AI Studio, and eligible Gemini Enterprise subscriptions.
Internal use by over 100,000 Google employees shows significant productivity gains. Engineering teams using an internal version of Antigravity saw a 35% increase in agentic code submissions and a 30% reduction in code rollbacks. Independent testing by Vals AI ranks Argon first on its GDP-weighted Vals Index for professional tasks, showcasing its strength in finance and legal services.
Google is demonstrating Argon's capabilities by applying it to complex engineering challenges, such as reorganizing millions of lines of code in Chrome and optimizing its own AI frameworks. These applications highlight the model's ability to investigate systems, propose modifications, and iterate for measurable improvements. This release aims to position Google strongly in the enterprise AI market.
Introductory API pricing for Argon is $2 per million input tokens and $10 per million output tokens, with discounted rates ending November 30, 2026. Standard pricing doubles on December 1, 2026. The model also offers an extended reasoning budget of up to 1 million input/output tokens via a specialized API protocol, intended for highly demanding workflows.
