CALIFORNIA / RankWire.AI / – Google has introduced Gemini 4 Argon, its latest flagship AI designed for complex professional tasks. The company announced this model on Sept. 30 as part of its Gemini 4 series. Argon is built to support software engineering, financial, legal work, and cybersecurity defense. Google claims the model can handle deeper reasoning over extended, multi-step processes. Currently, access is limited to a select group of cybersecurity defenders through its Fairwind Program.
Gemini 4 Argon increases the maximum token output to 1 million, up from the previous 64,000 tokens. This larger limit enables the model to process longer tasks within a single session. The initial API pricing is set at $2 per million input tokens and $10 per million output tokens. Input tokens stored in cache benefit from a 95% discount. After the launch period, prices are expected to rise to $4 and $20, respectively.
Google reports that thousands of its employees are already using Argon for coding, research, and writing projects. Internal teams are also applying the model to optimize data center memory and facilitate large code migrations. One initiative involved Argon agents migrating code from C and C++ to Rust. Another project used agents to profile memory across Google data centers. These efforts resulted in freeing more than 300 tebibytes of memory, with additional savings found through the work.
Enhanced Ability to Tackle Complex Professional Tasks
Google reports a 77.9% score for Gemini 4 Argon on DeepSWE v1.1, which evaluates long-term software engineering performance. The company also highlighted benchmarks in finance, legal work, and automation. Argon supports multimodal reasoning along with coding and enterprise activities. Developed by Google DeepMind, the model is part of the broader Gemini family. The increased output capacity helps Argon manage workflows that involve many steps of reasoning and execution.
Cybersecurity is another key focus for the model’s initial deployment. Google states Argon can identify, verify, and patch software vulnerabilities within controlled defensive environments. Wiz is leveraging Argon through its Scan for Good initiative, aimed at detecting security risks in public infrastructure. Google reports Argon scored 68% on CWE-bench v1, a vulnerability remediation benchmark. Selected defenders are granted access without the usual cyber safeguards for specific defensive tasks.
Limited Launch Precedes a Wider Release of Gemini 4
Google has not set a specific date for the general public release of Gemini 4 Argon. The company is employing a phased approach and gathering feedback from early testers. It is also part of a voluntary U.S. government process for pre-release model access. Future availability will include developers, enterprises, and consumers. The rollout begins with paid API users and Google AI Ultra subscribers, though no fixed launch date has been announced for these groups.
Additionally, Google has stated it will not release Gemini 3.5 Pro, which was initially scheduled for June. As a result, Gemini 4 Argon remains the company’s latest flagship AI for demanding reasoning and professional workloads. Other Gemini models are still available for different performance and cost needs. Argon introduces higher output, advanced coding features, and cybersecurity capabilities. Its current focus remains on trusted testers and select security partners.