Gemini 4 Argon brings advanced reasoning and a 1M token limit

Google announced the release of Gemini 4 Argon, a new frontier model built to handle complex and long-horizon professional workflows. The system is designed to support advanced reasoning across real-world software engineering, financial research, legal drafting, and cybersecurity defense.

Core Capabilities

Gemini 4 Argon features an industry-leading 1 million token limit, a significant increase from the previous 64K token limit. This expanded capacity allows the model to process large amounts of information and maintain deep reasoning across multi-step tasks. The model is initially rolling out to trusted cyber defenders through the Fairwind Program, with broader availability planned for developers, enterprises, and consumers following rigorous safety evaluations.

Performance in Professional Tasks

Google reports that Argon excels across various professional benchmarks. The model achieves a score of 77.9 percent on DeepSWE v1.1, which measures performance in real-world software engineering. It also leads the Vals Index for economic impact across finance, legal, and tax work, and ranks first on AutomationBench with a score of 51.3 percent.

Pricing and Availability

Argon launches with an introductory pricing structure of $2 per million input tokens and $10 per million output tokens. Cached input tokens are priced at a 95 percent discount compared to standard input tokens. Initial access is restricted to the Fairwind Program and internal Google teams, with wider release to paid API customers and Google AI Ultra subscribers scheduled after ongoing safety iterations.