Google introduced Gemini 4 Argon, its latest frontier artificial intelligence model, on September 30, 2026. The company stated that Argon is designed for "deep reasoning across complex, long-horizon workflows" and is set to deliver advanced performance in software engineering, enterprise knowledge work, and cybersecurity defense.
Initially, access to Gemini 4 Argon will be limited to a select group of "trusted cyber defenders" via Google's Fairwind Program. This phased release approach is intended to ensure safety and allow for rigorous testing before a broader public rollout. Google is actively participating in the U.S. government's voluntary process for pre-release model access. The company plans to gradually expand access to developers, enterprises, and consumers, starting with paid API customers and Google AI Ultra subscribers.
Gemini 4 Argon features an industry-leading 1 million token limit, enabling it to handle deep, multi-step problem-solving. Google claims the model sets new benchmarks in areas such as real-world long-horizon software engineering tasks, with a reported score of 77.9% on the DeepSWE v1.1 benchmark. It also demonstrates capabilities in autonomously finding, validating, and patching critical software vulnerabilities. For these cybersecurity defense functions, Google intends to release Argon without cyber guardrails to trusted defenders and its own internal teams.
The model's pricing structure includes an introductory offer of $2 per million input tokens and $10 per million output tokens, with cached input tokens at a 95% discount. After an introductory period, these prices are set to increase to $4 per million input tokens and $20 per million output tokens.
Internally, Google engineers have been using Gemini 4 Argon for various tasks, including debugging, large-scale codebase migrations, and algorithm design. One reported internal use case involved Argon agents identifying and applying memory optimizations across Google's data centers, which is estimated to save between 500 TiB and 1 PiB of memory.
Koray Kavukcuoglu, SVP of Google DeepMind and Chief AI Architect, emphasized that safely releasing such advanced capabilities requires a measured approach. He noted that Google is continuously gathering feedback from early testers to refine the model's safeguards.
Google's announcement positions Gemini 4 Argon as a response to increased competition in the frontier AI space from companies like OpenAI and Anthropic. The company aims to reassert its position in high-stakes AI applications, particularly in software development, professional knowledge work, and cybersecurity operations.