New Delhi: Google DeepMind announced Gemini 4 Argon a few weeks ago, initially restricting access to cybersecurity defenders. However, the rollout has now expanded, allowing paying subscribers to access the model through the Gemini app.
Gemini 4 Argon is now available globally to enterprise customers and Google AI Ultra and Pro subscribers through the Gemini app. At its introduction, Google detailed how it had used Gemini 4 internally to assist with software engineering, cybersecurity and research. The company also highlighted how the model had helped optimise Python codebases.
The AI model also contributed to the development of physical AI applications, including Google’s robotics research. It helped ground a humanoid robot in real-world physics and enabled it to better understand its surroundings.
With Argon, Google is focusing on practical and abstract applications of AI in mathematics and science, including robotics and nuclear fusion research.
Gemini 4 Argon joins the ranks of frontier models such as OpenAI’s GPT-6 Astra and Meta’s new Muse Wonder. Google has also said that, during development, it participated in the White House’s voluntary review process for new models, with Gemini 4 passing the assessment under its safety framework.
Google said Gemini 4 Argon is designed to refuse harmful requests related to cyberattacks and chemical, biological, radiological and nuclear (CBRN) threats while supporting legitimate scientific research.
The company has strengthened its safeguards for the rollout, including techniques to monitor the model’s internal activations for signs of misuse. These safeguards were tested by internal and external red teams using manual and automated attack methods.
