Will Gemini close the gap? Google launches its most powerful AI model for subscribers

Google has launched Gemini 4 Argon, its most advanced artificial intelligence model, available to paid subscribers. The model is designed for complex tasks involving extended reasoning and extensive data analysis, with an output ceiling of one million tokens. The starting price for Google AI Pro subscribers is $74.90 per month. The launch comes after Google restricted access to advanced models for free users.

Google launched on Tuesday evening Gemini 4 Argon, its most advanced artificial intelligence model to date, and opened it for use by the general public, developers, and business customers worldwide. The model, developed by the DeepMind division, is the first in the Gemini 4 family and is designed to handle complex tasks requiring extended reasoning, analysis of large amounts of data, and performing multi-step operations. Access is limited to paid subscribers: Google AI Pro subscribers ($74.90 per month) and Google AI Ultra subscribers (starting at $390 per month). Users of the free version and Google AI Plus subscribers do not receive access. The launch comes a week after Google announced restrictions on access to advanced models for free and Plus users. The model offers an output ceiling of one million tokens, significantly higher than competitors OpenAI and Anthropic, which limit to 128,000 tokens. Google claims the organization ranks first on the Vals Index, but a Bloomberg report cited skepticism among Google employees regarding its performance on everyday programming tasks. Google dismissed these claims. The launch price for the API is $2 per million input tokens and $10 per million output tokens, which will double after November 30.