Google Gemini 4 Argon's Introductory Deal Ends Soon: API Prices Double in December
Google's Gemini 4 Argon AI model will see its API prices double on December 1, according to the company's official documentation. Input token costs will rise from $2 to $4 per million, and output tokens from $10 to $20 per million, ending an introductory pricing period for developers.
The impending price increase means Gemini 4 Argon's standard API rates will match those listed for Anthropic's Claude Opus 5.5, while OpenAI's GPT-6 Astra remains more expensive per token. This adjustment shifts Argon from being cheaper than Astra at introductory rates to being competitive with Opus 5.5 in terms of raw token cost.
Token rates alone do not fully reflect the total cost of completing a task. Benchmarking by Artificial Analysis suggests that Argon's estimated cost per task will rise from $1.99 to approximately $3.98 after the price change, assuming consistent token usage. This contrasts with GPT-6 Astra's estimated $3.26 per task, indicating that Argon could become more expensive for certain workloads, even with its 95% discount on cached input tokens.
The article emphasizes that actual costs depend on various factors beyond token prices, including task complexity, prompt design, reasoning settings, and cache usage. Developers are advised to measure the cost per completed task within their specific workloads rather than relying solely on advertised API prices for comparison. Argon is positioned for complex software engineering, enterprise knowledge work, and cybersecurity defense, but its performance varies across benchmarks.
Google's internal results suggest productivity gains from Argon, with engineers reporting increased agentic code submissions by more than 35% and reduced rollbacks by nearly a third. These are company-reported internal results. Developers should evaluate Argon against their own use cases, considering both cost and its documented limitations and safety measures.
Share this article
Related reading
6 stories
AI could cause 'major' incidents within 2 years without safeguards, says Google DeepMind chair
Google DeepMind's chair warns AI could cause major incidents, a timely caution as Google's Gemini model prices rise.

SEA’s E-commerce Market to Reach US$400B by 2031: Xendit Report

Singapore Steps Up AI Safety With New Testing Guidelines, Taskforce

Figma announces new features to define how agents work on its platform

How Do Authorised Scams Hide Behind Legitimate Bank Transfers?

