Google announced this Wednesday, the 30th, Gemini 4 Argon, a new artificial intelligence (AI) model that takes the lead in the Gemini 4 line. According to the company, this is its most sophisticated model to date, designed to manage complex and long-duration activities in sectors such as programming, cybersecurity, and various professional functions.
This launch follows months of anticipation for the successor to previous generations of advanced Google models. Initially, Argon will be made available to a restricted group of cybersecurity partners, while the company itself conducts security assessments before any expansion of access. Google has not yet specified a date for the public availability of the model.
The company reported that Gemini 4 Argon set a new record in practical software development tasks and achieved first place in cybersecurity evaluations. Furthermore, the model demonstrated the best performance in another test focused on corporate activities, covering the financial and legal areas.
The use of the new model is not limited to external users. Google informed that Argon is already being used internally to optimize memory consumption in its data centers. This internal application allowed for the release of hundreds of terabytes of memory without the need for new hardware acquisition.
More information about the launch
Before releasing Gemini 4 Argon to a wider audience, Google opted to subject it to rigorous tests and security evaluations. The initial distribution will be made to selected partners in the field of cybersecurity. Additionally, the company is part of a voluntary program by the United States government to gain early access to AI models before their official launch.
Google communicated that it is conducting a pre-launch security assessment with the U.S. government. The company also plans to strengthen defenses against various forms of misuse, including known attacks like 'prompt injection', before increasing access to the model. This focus on security gains relevance at a time of growing scrutiny over the risks inherent in autonomous execution AI systems.
The emergence of Argon comes after a long period without the introduction of a top-tier model by Google. Previously, the company had worked on a version called Gemini 3.5 Pro but abandoned the project before its launch. Reuters reported that the development exceeded the planned timeline and that the model lost competitiveness against competing systems.
With Gemini 4 Argon, Google reaffirms its strategic focus on developing a high-performance model capable of competing directly with the most advanced systems of other AI corporations. The arrival of this new model coincides with the day after the signing of a voluntary AI security agreement by Google's CEO, Sundar Pichai, with the President of the United States, Donald Trump, during a meeting with executives from major technology companies held at the White House.
