# Google DeepMind releases Gemini 4 Argon to cyber defenders first

> Google DeepMind released Gemini 4 Argon, its first new frontier generation since Gemini 3, to cyber defenders first.

*The first new frontier model generation since Gemini 3 reaches paid API and AI Ultra users after an initial release to defenders in the Fairwind Program.*

By Behzad Hosseini · WireRead
Canonical: https://wireread.com/news/google-deepmind-releases-gemini-4-argon-to-cyber-defenders-first

Google DeepMind announced Gemini 4 Argon on 30 September 2026, its first new frontier model generation since Gemini 3, aimed at long-horizon coding, legal and finance work, and cyber defence.

Google said in its [launch post](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/) that access begins with trusted cyber defenders in its Fairwind Program, with US government pre-release access. Paid API users and Google AI Ultra subscribers follow. Broader developer, enterprise and consumer release is "coming soon", the company said.

Introductory pricing is $2 per million input tokens and $10 per million output tokens, with cached input 95% cheaper. After the introductory period, prices rise to $4 per million input tokens and $20 per million output tokens.

The output limit rises to 1 million tokens, from 64K. Google describes this as industry-leading.

Google reported benchmark results of its own. Argon scored 77.9% on DeepSWE v1.1 and 68% on CWE-bench v1, a vulnerability remediation test, where it tied for first. It ranked first on AutomationBench at 51.3% and scored 91.7% on LVBench, a long-video test. Google said it leads the Vals Index across finance, coding, legal and tax, and leads Gray Swan's prompt-injection robustness test.

Google positions Argon against OpenAI's GPT-6 Astra and Anthropic's Claude Opus 5.5. Koray Kavukcuoglu, Google DeepMind's SVP and Chief AI Architect, wrote in the launch post: "Argon is helping our quantum computing researchers optimize the spacetime resources (qubits × gates)".

CNBC [asked whether Argon can catch up](https://www.cnbc.com/2026/10/02/tech-download-google-argon-frontier-openai-anthropic.html) with OpenAI and Anthropic at the frontier. One analyst told the outlet that "the final proof will be in enterprise production environments once the model is fully released".

Bloomberg reported that some Google employees privately doubt the model's real-world coding ability. Google has not confirmed this. The broader developer, enterprise and consumer release has no date beyond "coming soon".

## Key takeaways

- Argon goes first to Fairwind Program defenders, then paid API and AI Ultra users.
- Introductory pricing is $2 per million input tokens and $10 per million output.
- Output limit rises to 1 million tokens from 64K.

## Sources

- [Gemini 4 Argon](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/) — Google, 2026-09-30
- [Can Google's new model really catch up to OpenAI and Anthropic at the frontier?](https://www.cnbc.com/2026/10/02/tech-download-google-argon-frontier-openai-anthropic.html) — CNBC, 2026-10-02