Google Launches Gemini 4 Argon With New API Pricing

Google announced the Gemini 4 Argon model on October 13, 2026. This model is considered the flagship for handling challenging, multi-step tasks in software engineering, finance, legal, cybersecurity and scientific fields. Gemini 4 Argon is available to developers and enterprises worldwide in the Gemini API, Google AI Studio and Google's Gemini platform.

Gemini 4 Argon Availability and Subscription Plans

This latest update provides developers with a powerful new tool for challenging tasks, but access and cost factors will vary depending on the product. Google AI Pro and Google AI Ultra subscribers can access Gemini 4 Argon through the Gemini app and AI Mode in Search, and Google AI Ultra subscribers can access it via the Antigravity coding platform.

Gemini 4 Argon's introductory API has a starting price of $2 per million input tokens and $10 per million output tokens, compared to $4 per million input tokens and $20 per million output tokens after November 30, 2026. Cached input tokens received a 95% discount on the applicable output price during the applicable period.

Key Features and Technical Capabilities of Gemini 4 Argon

The Gemini 4 Argon model accepts text, images, videos, audio files, and PDFs as input and has a total context window of 1,048,576 tokens for text, including inputs and outputs. In addition, its long-decode capability allows for responses up to one million tokens in length. Among the available features are code execution, function calling, search grounding, file search and computer use in preview mode.

Gemini 4 Argon has the highest score on the Vals Index, a proprietary industry benchmark that includes financial, coding, legal, and tax-related tasks, according to Google. Gemini 4 Argon had a 68.9% score on the model card published by Google. These are the claims of the company as a general indicator of the capabilities of the model, but not an objective assessment of performance on any particular task.

Benchmark Results, Real-World Performance, and Pricing Outlook

Google also reports that its own productivity with Gemini 4 Argon has increased: for example, the number of agentic code submissions rose by more than 35%, and the number of code rollbacks decreased by nearly one-third. It remains to be seen whether independent developers will be able to achieve the same performance. Pay attention to the emergence of independent performance evaluations and changes in the dynamics of use in the future, as the introductory API pricing for Gemini 4 Argon is about to expire.

#artificial intelligence #google #gemini #ai models #gemini4 #argon