Google launches Gemini 4 Argon with introductory pricing that doubles in November

Google has officially launched Gemini 4 Argon, a new artificial intelligence model designed for complex enterprise and software development tasks. According to InfoWorld, the company positions the system as a flagship option for long-horizon agentic work, offering advanced reasoning capabilities alongside an introductory pricing structure that is scheduled to double after November 30.

Koray Kavukcuoglu, senior vice president of Google DeepMind and chief AI architect at Google, stated in a blog post that the model excels at deep work in complex, real-world tasks. In a separate developer update, Google described the system as being engineered for workflows that demand sustained reasoning and structural precision over extended periods.

Technical capabilities and internal testing

Google shared specific performance metrics from internal engineering efforts to demonstrate the model’s utility. The company reported that its engineers utilized Argon to refactor a legacy folder containing approximately six million lines of code within Chromium’s 40 million-line codebase. According to Computerworld, this massive refactoring effort was completed in weeks, a timeline that would have previously taken months.

Teams operating within Google’s Antigravity agentic development environment also reported significant efficiency gains. The company said these teams increased their agentic code submissions by more than 35 percent and reduced code rollbacks by nearly a third. Google did not provide specific methodologies for how it measured these submission and rollback figures.

Pricing dynamics and analyst warnings

The financial structure of the launch has drawn attention from industry analysts. During the introductory period, Argon costs $2 per million input tokens and $10 per million output tokens, with cached inputs receiving a 95 percent discount. A footnote in the launch materials indicates that these rates will rise to $4 and $20, respectively, after November 30. The one-million-token limit in the Gemini API is shared between input and output.

Sanchit Vir Gogia, chief analyst at Greyhound Research, advised CIOs to budget at the higher standard rate from the start. He noted that enterprise evaluation, security review, and procurement cycles rarely finish within the short discount window. Gogia warned that a business case built on launch rates is likely to require rewriting within seven weeks.

Gogia added that Google bills reasoning tokens at the output rate, which complicates cost estimation for agentic workloads. He argued that cheap per-token pricing does not necessarily equate to cheap per-task costs, and only the latter impacts the actual budget.

Market context and adoption risks

Sakshi Grover, research director at IDC, highlighted the inherent cost risks of agentic workloads. She explained that these systems multiply token usage through planning loops, tool calls, and repeated context management. Grover cited IDC’s FERS Survey Wave 5 from 2026, which found that 72 percent of enterprises reported their AI agent spending exceeded forecasts in the previous 12 months.

Argon is accessible through the Gemini API in AI Studio, on the Google Gemini platform, and to Google AI Ultra subscribers via the Antigravity environment. This launch follows earlier announcements by Google regarding Gemini 4 Argon’s capabilities, including its introduction with million-token outputs as Threat Times reported on October 14, 2026, and its application in robotics control as noted on October 13, 2026.

What should you do?

Security and technology professionals should review their current AI spending forecasts to account for potential price increases. Organizations using agentic workflows should monitor token consumption closely, particularly for reasoning tokens that are billed at higher output rates. Teams should evaluate whether the introductory pricing provides genuine savings or merely delays standard costs.

What to do and how to stay safe: Google

• Audit current AI agent token usage to estimate costs at the post-November 30 price levels.
• Review procurement timelines to determine if any projects can be completed before the price increase.
• Implement strict token limits and monitoring for agentic workflows to prevent runaway spending.
• Evaluate alternative models or cached input strategies to mitigate potential cost overruns.

Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

When do the prices for Gemini 4 Argon increase?

The introductory pricing for Gemini 4 Argon is set to double after November 30, rising to $4 per million input tokens and $20 per million output tokens.

How did Google test the model's coding abilities?

Google engineers used Argon to refactor six million lines of code in Chromium, completing the task in weeks instead of months, according to company statements.

What do analysts say about budgeting for this model?

Analysts recommend budgeting at the higher post-discount rate immediately, as enterprise procurement cycles typically exceed the short introductory pricing window.