CALIFORNIA / RankWire.AI / – Google has launched Gemini 4 Argon, its newest leading artificial intelligence model designed for complex professional tasks. Announced on Sept. 30, Argon stands at the forefront of the Gemini 4 series, enabling advanced functions such as sophisticated software development, financial analysis, legal research, and cybersecurity operations. It is also capable of managing lengthy reasoning and execution sequences. Currently, early access remains restricted, with selected cybersecurity specialists utilizing Argon through the Fairwind Program.
Argon introduces a maximum token output of 1 million, a significant jump from the former limit of 64,000 tokens. This enhancement allows the model to perform extended tasks without needing to split work across multiple sessions. Initial API pricing begins at $2 per million input tokens, with output tokens costing $10 per million during the same period. Inputs stored in cache benefit from a 95% discount. Future pricing adjustments are expected to be $4 for input tokens and $20 for output tokens.
Currently, thousands of employees at Google are leveraging Argon for tasks including coding, research, and content creation. Internal testing has taken place on projects involving data center optimization and large-scale software migrations. One such project involved deploying Argon agents to facilitate C and C++ to Rust transitions, while another focused on memory profiling across data centers. These efforts resulted in freeing over 300 tebibytes of memory, with teams identifying further efficiencies through ongoing system analysis.
Enhanced capacity for extensive technical tasks
Google reported a 77.9% score for Argon on DeepSWE v1.1, an assessment of long-term software engineering performance. The company also shared results in areas such as finance, legal tasks, automation, and multimodal functionalities. Argon was developed by Google DeepMind as part of the broader Gemini model family, which integrates coding utilities with long-context reasoning and multimodal capabilities. Its expanded output ability enables the completion of workflows that involve multiple interconnected steps before reaching the final outcome.
Cybersecurity remains a key focus during the initial rollout. Argon is capable of detecting, verifying, and patching software vulnerabilities within approved defensive environments. Wiz is utilizing the model through its Scan for Good initiative, targeting security flaws in public infrastructure. Google reported a 68% score on CWE-bench v1, a benchmark designed to evaluate vulnerability remediation skills. Selected cybersecurity defenders can also operate Argon outside standard safeguards when working on authorized security tasks.
Limited public access as rollout progresses in stages
Google has not announced a specific date for widespread public access to Gemini 4 Argon. The deployment is being conducted gradually, with feedback being collected from early users. The company is also involved in a voluntary U.S. government process that grants pre-release access to advanced AI models. Broader availability is expected for developers, enterprise clients, and consumers over time. Priority access will likely be given to paid API subscribers and Google AI Ultra members, although no official launch date has been disclosed.
The firm also clarified that Gemini 3.5 Pro will not be released. Although it was anticipated before Gemini 4, Argon now serves as the flagship model for demanding reasoning and professional applications. Other Gemini models continue to be available for users with different performance and cost requirements. For now, Gemini 4 Argon remains focused on trusted testers, cybersecurity collaborators, and limited early-access programs, with wider distribution still pending.