CALIFORNIA / RankWire.AI / – Google has launched Gemini 4 Argon, its flagship artificial intelligence system designed for sophisticated professional applications. Announced on Sept. 30, Argon is positioned as the leading model within the Gemini 4 series, capable of supporting tasks such as advanced software development, financial analysis, legal research, and cybersecurity operations. The model is also equipped to manage extended sequences of reasoning and execution. Currently, access to Argon is limited to select cybersecurity defenders participating through the Fairwind Program.
The maximum token output for Argon has been increased to 1 million, a significant jump from the previous limit of 64,000 tokens. This enhancement enables the model to perform longer and more complex tasks without requiring work to be split into multiple sessions. Initial API pricing begins at $2 per million input tokens, while output tokens are charged at $10 per million during the same period. Inputs stored in cache benefit from a 95% discount. Future pricing adjustments are expected to raise costs to $4 for input tokens and $20 for output tokens.
A considerable number of employees within the company already utilize Argon for activities such as coding, research, and writing assignments. Internal teams have tested the model on optimizing data centers and migrating large-scale software systems. One project employed Argon agents to facilitate C and C++ migrations to Rust, while another focused on memory profiling across data centers. These efforts led to freeing more than 300 tebibytes of memory and uncovered additional savings through ongoing analysis of the same systems.
Enhanced capacity for long-form technical tasks
Google reported a 77.9% score for Argon on DeepSWE v1.1, a benchmark for evaluating extended software engineering performance. The company also published results related to finance, legal tasks, automation, and multimodal operations. Developed by Google DeepMind as part of the broader Gemini model family, Argon merges coding capabilities with long-context reasoning and multimodal functions. Its increased output capacity is tailored to support projects requiring multiple interconnected steps before completion.
Cybersecurity remains a primary focus during this initial rollout. Argon can detect, validate, and remediate software vulnerabilities within authorized defense environments. Through its Scan for Good initiative, Wiz is utilizing the model to identify security flaws in public infrastructure. Google’s reports indicate a 68% score on CWE-bench v1, a benchmark focused on vulnerability remediation. Selected cybersecurity teams can also access Argon without standard safety measures when working on approved security tasks.
Limited public access during phased deployment
Google has not announced a specific date for widespread public availability of Gemini 4 Argon. Instead, the company is rolling out the model gradually and collecting feedback from early users. It is also participating in a voluntary process with the U.S. government that grants pre-release access to advanced AI systems. Eventually, broader access will include developers, enterprise clients, and consumers. Priority will be given to paid API subscribers and Google AI Ultra members, although no official launch date has been fixed.
Additionally, Google confirmed that Gemini 3.5 Pro will not be released. This model was previously anticipated prior to the debut of Gemini 4. Argon now stands as the flagship AI offering for complex reasoning and professional workloads. Other Gemini models are still available, catering to users with varying performance requirements and budgets. For the time being, Gemini 4 Argon remains limited to trusted testers, cybersecurity partners, and select early-access programs as the broader rollout is yet to commence.