Google Unveils Gemini 4 Argon, Its Latest AI Model, for Cybersecurity Defenders

Google has unveiled Gemini 4 Argon, its latest flagship AI model. The company has given it to a small group of trusted cybersecurity defenders through its Fairwind Program. These users are getting access to the full abilities of Argon without its cyber guardrails.

The model is designed for long and complex workflows in software engineering, legal work, finance, and cybersecurity defense. It will also be available to paid API customers and Google AI Ultra subscribers after testing. The company claims that Argon sets a new high on DeepSWE v1.1 with 77.9% performance and ties for first place on CWE-bench with 68%.

Google is also raising the model's output limit to 1 million tokens, up from 64,000. The introductory price will be $2 per million input tokens and $10 per million output tokens. However, some Google employees are skeptical about the performance of Argon in certain areas.