CALIFORNIA / RankWire.AI / – Google has introduced Gemini 4 Argon, its latest premier artificial intelligence system designed for handling intricate professional tasks. The company announced this model on Sept. 30, marking it as the cornerstone of the Gemini 4 series. Argon is tailored for software engineering, financial, legal expertise, and cybersecurity defense. Google claims the model is capable of sustaining more in-depth reasoning over extended, multi-step processes. Currently, access is being granted to a select group of cybersecurity defenders through its Fairwind Program.
With Gemini 4 Argon, the maximum output limit has been increased to 1 million tokens, a significant rise from the previous 64,000 tokens. This larger capacity enables the model to process long-duration tasks within a single processing run. Google has set the initial API pricing at $2 per million input tokens and $10 per million output tokens. Input tokens stored in cache benefit from a 95% discount off the input price. After the introductory phase, prices are expected to increase to $4 and $20, respectively.
Google reports that thousands of its employees are already utilizing Argon across specialized programming, research, and writing projects. Internal teams have also employed the model for data center memory optimization and extensive code migrations. One initiative involved using Argon agents to transition C and C++ code to Rust. Another focused on memory profiling across Google’s data centers. The company states that these optimizations have freed over 300 tebibytes of memory, with additional savings identified through ongoing work.
Enhanced Capacity for Complex Professional Tasks
Google reported a 77.9% score for Gemini 4 Argon on DeepSWE v1.1, a benchmark assessing long-term software engineering performance. The results also include performance metrics in finance, legal work, and automation benchmarks. Argon is capable of multimodal reasoning alongside coding and enterprise applications. Developed by Google DeepMind within the broader Gemini family, the model’s increased output capacity helps Argon manage extended workflows involving multiple consecutive reasoning and execution stages.
A significant aspect of the model’s initial deployment is its cybersecurity capabilities. Google indicates Argon can detect, validate, and patch software vulnerabilities in controlled defensive environments. Through its Scan for Good initiative, Wiz employs Argon to identify security vulnerabilities in public infrastructure. Google reports that Argon achieved a score of 68% on CWE-bench v1, a benchmark for vulnerability remediation. Selected cybersecurity defenders are granted access to the model without the usual safety protocols to carry out authorized defensive tasks.
Initial Limited Deployment Prepares for Broader Gemini 4 Access
Google has not yet announced a specific release date for the wider public availability of Gemini 4 Argon. The company is adopting a phased rollout approach and is gathering feedback from early testers. It also participates in a voluntary U.S. government process for pre-release model access. Future plans include making the model accessible to developers, enterprises, and consumers. The initial rollout will target paid API customers and Google AI Ultra subscribers, although no fixed launch date has been provided for these groups.
Additionally, Google has stated it does not plan to release Gemini 3.5 Pro, which was originally scheduled for June. As a result, Gemini 4 Argon remains the company’s latest flagship model for demanding reasoning and professional applications. Google continues to offer other Gemini variants to suit different performance levels and budget needs. Argon distinguishes itself with its larger output, enhanced coding features, and dedicated cybersecurity functions. Its current availability remains limited to trusted testers and selected security partners involved in defense applications.